Granuloma annulare: review of cases in the Social Hygiene

Transcription

Granuloma annulare: review of cases in the Social
Hygiene Service and a case-control study of its
association with diabetes mellitus
Dissertation for exit assessment of higher physician training,
Specialty in Dermatology and Venereology,
Hong Kong College of Physicians
Chan Po Tak
December 2003
Contents
List of tables
List of figures
List of colour plates
Abbreviations used in the text
Abstract
Chapter 1 Introduction
iv
v
vi
vii
viii
1
Chapter 2 Methodology
A) Identification and definition of GA cases
B) Interview with GA cases
C) Histopathology slide review
D) Recruitment of controls
E) Choice of FBS as a measure for DM
1) Reasons for using FBS
2) Limitations of using FBS alone
3) Problem of diagnostic bias
4) Reasons for not using other sensitive measures for
carbohydrate intolerance
E) Statistics
4
4
5
6
6
6
7
7
7
8
Chapter 3 Results
A) Demographic data
B) Clinical features
1) Classification of subtypes of GA
2) Distribution
3) Morphology of individual lesion
4) Number of skin lesions
5) Symptoms
6) Precipitating factors
C) Pathological features
D) Concurrent diseases
E) Associated DM and previous DM screen in skin clinics
F) Case-control study of association with DM
G) Family history of GA and DM
H) Duration of follow up
I) Treatment
J) Prognosis
9
10
10
12
13
13
13
14
15
15
16
18
19
19
19
21
Chapter 4: Discussion
A) Epidemiology
24
25
i
1) Morphology of skin lesions
2) Classification
3) Other variants of GA reported in the literature
4) Distribution of skin lesions
5) Symptoms
C) Pathology
1) Necrobiosis and mucin deposition
2) Inflammatory cells
3) Elastic fibre changes
4) Vascular changes
5) Subcutaneous and perforating GA
D) Etiology
1) Genetic Factors
2) Physical Factors
3) Diseases associations
4) Association with DM
5) Drugs
E) Pathogenesis
1) Cell mediated immunity
2) Immune complex vasculitis
3) Collagen abnormalities
F) Differential diagnosis
G) Treatment
1) Resolution after skin biopsy
2) Steroids
3) Cryotherapy
4) Intralesional interferon
5) Surgery
6) Topical imiquimod
7) Cream psoralen plus ultraviolet A
8) Photo(chemo)therapy
9) Isotretinoin
10) Dapsone
11) Other second line therapy for generalized GA
H) Prognosis
I) Limitations of the present study
J) Conclusions and suggested further study
25
25
26
27
27
28
28
28
28
29
29
29
30
30
30
30
32
33
33
33
34
34
34
35
35
36
36
36
36
36
36
37
37
38
38
39
39
40
ii
References
Acknowledgement
Appendix
1. Data sheet for GA cases
2. Data sheet for control
3. Memo to GA cases
4. Memo to control
5. Definitions of terms
Colour plates
41
49
50
51
52
53
54
56
iii
List of tables
Table 1. Summary of published articles on the relationship between GA and DM
2,3
Table 2. Diagnostic criteria of GA employed in the present study
5
Table 3. Total number of GA cases identified and recruited into this study
4
Table 4. Exclusion criteria for controls
6
Table 5. Values for diagnosis of DM and impaired fasting glucose
7
Table 6. Age of onset of GA in relationship with gender
9
Table 7. Classification of subtypes of GA in this study
10
Table 8. Distribution of skin lesions in various subtypes of GA
12
Table 9. Morphology of GA lesions
13
Table 10. Reported precipitating factors for GA
14
Table 11. Histology findings in GA
15
Table 12. Concurrent medical illnesses reported by GA cases
16
Table 13. Previous screen for DM and results
16
Table 14. Distribution of DM in cases (at first FBS screen) and controls
18
Table 15. Response to treatment in GA cases
19
Table 16. Summary of treatment modalities used in GA cases
20
Table 17. Age and sex distribution of GA in the literature
24
Table 18. Subtypes of GA in various case series
26
Table 19. Proportion of symptomatic GA cases in the literature
28
Table 20. Pattern of histiocytic infiltration in GA
28
Table 21. Reported diseases associated with GA
30,31
Table 22. Therapeutics for GA
35
iv
List of figures
Figure 1. Age of onset of GA in different sex
9
Figure 2. Sex distribution of GA cases
9
Figure 3. Distribution of various GA subtypes
11
Figure 4. Age of onset of GA in localized and generalized subtypes
11
Figure 5. Proportion of symptomatic skin lesions in various GA subtypes
14
Figure 6. Type of glucose intolerance in GA cases
17
Figure 7. Proportion of first CR of all GA cases with time
21
Figure 8. Proportion of first CR in localized and generalized GA
21
Figure 9. Effect of diabetic status on first CR of GA
22
Figure 10. Effect of age of onset on first CR of GA cases
22
Figure 11. Effect of sex on first CR of GA cases
23
v
List of colour plates
Plate 1. Localized GA with skin coloured papules over the left elbow
56
Plate 2. Typical GA with erythematous nonscaly arciform lesions over the left
elbow
56
Plate 3. GA with annular lesions over both hand dorsa and papular lesions over the
distal forearms
Plate 4.GA: predominant papular form with slightly erythematous papules over
the distal forearms
Plate 5. Generalized GA: coalescing papules without central clearing, forming
plaques over both elbows
Plate 6. Generalized GA with well defined erythematous plaques over the
abdominal wall
Plate 7 a and b. Generalized GA with extensive coalescing papules over the
chest, abdomen, back and arms
Plate 8. Generalized GA with annular lesions over the abdominal wall
57
Plate 9. Generalized GA with multiple lesions over the right elbow
60
Plate 10. Generalized GA, same patient as Plate 8
61
Plate 11. Generalized GA with annular and serpiginous lesion over the right
abdominal wall and right hip
Plate 12. Perforating GA with skin colour papules over the hand dorsum
61
Plate 13. Histology: typical palisading granuloma with central necrobiosis
62
Plate 14. Histology: interstitial pattern showing histiocytes
63
Plate 15. Histology: histiocytes and occasional giant cells are the main
inflammatory cells in GA
Plate 16. Histology: nuclear dust can sometimes be identified in GA lesions
63
64
Plate 17. Histology: perforating GA with a ‘busy looking’ superficial dermis
64
Plate 18. Histology: perforating GA (higher power of Plate 17)
65
57
58
58
59
60
62
vi
Abbreviations used in text
CI
CR
DM
FBS
HLA
IFG
IFN-γ
GA
OGTT
PUVA
SD
TIMP
Confidence interval
Complete remission
Diabetes mellitus
Fasting blood sugar
Human leucocyte antigen
Impaired fasting glucose
Interferon-γ
Granuloma annulare
Oral glucose tolerance test
Psoralen plus ultraviolet A
Standard deviation
Tissue inhibitor of metalloproteinases
vii
Abstract
The study included thirty-one granuloma annulare (GA) cases. Male to female
ratio was 1.82:1. Age of onset was distributed bimodally in both sexes, with a median
age of onset of 45 in male and 61 in female. 38.7% had onset of disease in the first three
decades. Localized, generalized, perforating and subcutaneous GA accounted for 67.8%,
25.8%, 3.2% and 3.2% of cases respectively. They were mostly found on distal
extremities (excluding palms and soles). A predominantly annular morphology was the
commonest (48.4%). In most cases, symptoms were absent (77.4%) and no precipitating
factors could be identified (90.3%). Over 80% of GA cases were followed up for more
than two years. Three localized GA resolved within one month after skin biopsy.
Complete remission (CR) was achieved in 17 GA cases by either topical steroid alone (16
cases) or together with intralesional steroid (one case). Another generalized GA had CR
after six months of oral isotretinoin. Dapsone was tried in one case but failed and CR
was finally achieved after topical and intralesional steroid. Seven cases had persistent
disease after topical steroid. The median time for first remission for localized and
generalized GA were 39 months and 54 months respectively, but the difference was not
statistically significant. Age, sex and DM status did not affect duration of disease.
33.3% of remitted cases had disease recurrence after a median time of 9.5 months.
Pre-existing diabetes mellitus (DM) was present in four cases before GA (median
three years), DM occurred after GA in three cases (median five years) and impaired
fasting glucose was detected 2.5 years after GA in one case. Twenty two GA cases, with
enough age and sex matched controls in a 1 to 3 ratio, were used in a case-control study
on the association of GA with DM and the result was not statistically significant.
Keywords: granuloma annulare, diabetes mellitus, case-control study
viii
Chapter 1: Introduction
Granuloma annulare (GA) is a degenerative disease of the skin, characterized by
focal degeneration of collagen with surrounding areas of reactive inflammation and
fibrosis.1 The lesion of GA was first described by Colcott Fox in 1895 and was named by
Radcliffe-Crocker in 1902. 2 There are several clinical manifestations, ranging from
localized GA, which is the most common form, to less commonly seen variants,
including generalized, subcutaneous and perforating GA.3,4,5
Etiology and pathogenesis of GA remains unknown and speculative. The
relationship between GA and diabetes mellitus (DM) has been extensively studied. There
is controversy over such an association, with some studies indicate a
relationship6,7,8,9,10,11,12,13 while others refute the linkage.14,15 Some studies have shown
relationship of DM with generalized GA,7,9 whereas others have detected its association
with localized GA10 (Table 1). These studies, albeit with different conclusions, are not
directly comparable to one another. Their differences can be broadly classified into the
following categories:
1) Choice of GA cases: Different studies have variable proportions of GA subtypes.
Moreover, not all GA cases are biopsy proven. Clear diagnostic criteria for GA
have never been given.
2) Choice of controls: Controls are preferably matched in terms of age, sex and race,
as these variables may affect the prevalence of DM. They should also be free of
skin diseases that are related to DM.
3) Choice of measures for carbohydrate intolerance: The different methods include
fasting blood sugar (FBS),7,9 oral glucose tolerance test (OGTT),7,8,11
haemoglobin A1c,14 cortisone-glucose tolerance test,6,7,8,9 insulin–glucose curve,8
intravenous glucose tolerance test,7,9,11 tolbutamide test9 and postal questionnaire
to attending general practitioner.10 As different diagnostic criteria have been used,
some authors prefer the term carbohydrate intolerance instead of DM.9
A retrospective study on GA has been carried out in the Social Hygiene Service of
the Department of Health. The service comprises clinics covering the major territories of
Hong Kong Special Administrative Region and is the major public centre receiving
referrals for dermatological problems and sexually transmitted disease. The aims of the
study are to review the clinical features of GA in our local population and to explore
whether there is any association of GA with DM by a case-control study.
1
Table 1. Summary of published articles on the relationship between GA and DM
Source
Case
Control
Measures of
diabetic state
Results
Rhodes EL,
6
et al 1966
30 nondiabetic
GA patients, not
mentioned
percentage of
localized GA
versus
generalized GA,
all were biopsy
proven
15 hospital patients
and staff of same age
group, no family
history of DM, recent
infection, surgery or
treatment with drugs
that might alter
carbohydrate
metabolism
50 gm OGTT,
prednisone glycosuria
test
10 GA patients showed
abnormal rise in glucose
excretion in the prednisone
glycosuria test, 2 of which
also had abnormal glucose
tolerance; whereas no
abnormal values were
obtained in controls of a
similar age group
Haim S, et al
7
1970
8 generalized
GA cases, all
were skin biopsy
proven
Nil
For nondiabetic cases,
FBS was determined.
For cases with normal
FBS, oral or
intravenous glucose
tolerance tests were
performed.
If above tests all
normal, prednisone
glucose tolerance test
was used
For the first patient, FBS was
normal, but other tests were
not carried out. 3 cases had
DM before GA. For the
remaining 4 cases, 2 had
abnormal glucose tolerance
test, another 2 had normal
glucose tolerance test but
pathological prednisone
glucose tolerance test
Hammond R,
8
et al 1972
10 nondiabetic
adults with GA:
9 localized, 1
generalized; all
were verified by
skin biopsy; age
22 to 28
Normal adults,
matching criteria and
number of controls
were not mentioned;
age 20-62; all were
subsequently normal
on both OGTT and
cortisone-glucose
tolerance test
OGTT; cortisoneglucose tolerance test;
Insulin to glucose ratio
OGTT: abnormal in 2 GA;
Cortisone-glucose tolerance
test: abnormal in another 3
GA, 1 GA borderline; Insulin
to glucose ratio: lowered
insulin response in remaining
4 GA; results suggested a
decreased pancreatic beta cell
sensitivity to glucose
stimulation in GA
Haim S, et al
9
1973
52 GA cases: 39
localized, 13
generalized, all
were proven by
skin biopsy
2 control groups:
1) Same number of
patients with other
skin diseases
unrelated to DM
2) 386 patients with
conjunctival
aneurysms in another
medical centre
FBS for known DM;
Prednisone-glucose
tolerance test for nondiabetics;
For borderline cases on
prednisone- glucose
tolerance test,
intravenous glucose
tolerance test and
insulinogenic reserve by
tolbutamide test were
also determined
Carbohydrate intolerance in
controls =23%, in conjunctival
aneurysms controls= 11.67%,
in generalized GA= 76.9%, in
localized GA=23.1%.
Incidence of carbohydrate
intolerance was similar to
controls in localized GA, but
was higher than controls in
generalized GA.
Muhlemann
MF, et al
10
1984
557 GA cases:
511 localized,
38 generalized,
8 with
insufficient
information for
classification;
only 10% was
skin biopsy
proven
Age-matched
population data on
diabetic prevalence,
one in Denmark, one
in England
Questionnaire to
patient’s general
practitioner on diabetic
status
24 cases of DM were
identified (18 type 1 DM, 6
type 2 DM); in all but one
case DM was associated with
localized GA; GA was
associated with type 1 DM, no
analysis was performed for
type 2 DM
2
Table 1 (continued). Summary of published articles on the relationship between GA
and DM
Source
Case
Control
Measures of
diabetic state
Results
Kidd GS, et
11
al 1985
21 GA cases, all
biopsy
confirmed, 3
disseminated, 1
perforating, 17
localized. Only
15 patients were
included in
formal study; 4
known DM
patients and 2
obese patients
were excluded
14 weight- and agedmatched controls, all
were healthy and not
on any medications
100 gm OGTT and 25
gm intravenous glucose
tolerance test
OGTT: fasting plasma
glucose, 2 hour glucose, 1
hour serum insulin, area
glucose curve and area insulin
curve were significantly
higher in GA patients.
Intravenous glucose tolerance
test: fasting plasma glucose
and area glucose curve were
higher in GA patients.
Results showed higher glucose
value in association with
higher insulin level,
suggesting insulin resistance
Shimizu H,
12
et al 1995
Literature
review of 23
perforating GA
cases; arbitrarily
divided into 2
types: papular
perforating type
and ulcerative
perforating type
Nil
Not stated in article,
just mentioned whether
DM was present or not
Overall 30% of reported
perforating GA had DM. For
those cases with DM, 86% had
ulcerative lesions and 14%
had papular perforating
lesions. Authors concluded
that DM played a role in the
ulcerative perforating GA
Veraldi S, et
13
al 1997
61 GA cases: 35
localized, 26
generalized, all
were proven by
skin biopsy
126 age- and sexmatched patients with
different
inflammatory skin
diseases admitted to
hospital
Plasma glucose, glucose
tolerance test
Overall DM rate in all GA
cases was not significantly
different from controls.
Subgroup analysis: type 1 DM
was associated with GA,
whereas type 2 DM was not
associated with GA.
Gannon TF,
14
et al 1994
23 GA cases: 13
localized, 10
generalized; all
were proven by
skin biopsy
No control was
recruited because all
GA cases had normal
fasting plasma
glucose and
haemoglobin A1c
Fasting plasma glucose
and haemoglobin A1c
Carbohydrate intolerance was
absent in localized and
generalized GA cases.
Nebesio CL,
15
et al 2002
126 GA cases
identified (only
18% biopsy
proven), cases
below 10 years
old were
excluded for
lack of control,
cases not seen
by dermatologist
were also
excluded. 50
GA cases were
finally recruited
50 psoriasis patients
who were matched
by age, sex and race
Blood glucose
measurements,
haemoglobin A1c,
classification of DM as
reported by care
providers
No Type 1DM identified in
GA cases. 22% of GA cases
had type 2 DM, which was not
significantly different from
controls (20%).
3
Chapter 2: Methodology
A) Identification and definition of GA cases
The study was carried out in the Social Hygiene Service and was approved by the
Ethics Committee of the Department of Health. The skin biopsy records of the following
clinics were scrutinized manually from 1st January 1990 to 31st July 2002 and results with
keywords ‘granuloma annulare’ were identified:
Chai Wan Social Hygiene Clinic
Cheung Sha Wan Dermatological Clinic
Lek Yuen Social Hygiene Clinic
Sai Ying Pun Dermatological Clinic
South Kwai Chung Social Hygiene Clinic
Tuen Mun Social Hygiene Clinic
Wan Chai Social Hygiene Clinic
Yau Ma Tei Dermatological Clinic
Yung Fung Shee Dermatological Clinic
The clinical records, pathology reports and clinical photographs were assessed to
see whether cases met the diagnostic criteria of GA in this study (Table 2, p.5). The
clinical study only included those cases meeting the diagnostic criteria of GA, had
retrievable clinical record and were able to attend a personal interview or telephone
interview with the author.
B) Interview with GA cases
Details of GA and response to treatment were recorded in the patient data sheet
during the interview (Appendix 1). Physical examination was performed for cases who
were personally interviewed. For cases not having DM, FBS was arranged. A total of 83
cases of GA were identified. Only 31 cases were included in the present study (table 3).
Table 3. Total number of GA cases identified and recruited into this study
GA identified by biopsy record
Record lost/ not retrievable
Histology suggestive of GA, but not confirmed clinically
Histology suggestive of GA, but rejected after subsequent
biopsies
Case rejected after assessment with pathologist because
histological criteria were not met
Failure to contact cases
Cases who refused to join the study
Cases personally interviewed
Cases interviewed over phone only
Number of cases
83
23
2
2
2
18
5
29
2
4
Table 2. Diagnostic criteria of GA employed in the present study
1.
Clinically in typical GA, skin lesions are characterized by both:
a)
skin-coloured, erythematous to violaceous dermal papules that may be
solitary or arranged in an arciform and circular pattern. In circular form, the
central surface may be slightly hyperpigmented;
and
b)
lack of epidermal change
or in case of atypical forms of GA:
i)
ii)
2.
Subcutaneous GA- skin coloured subcutaneous nodule that may be solitary
or multiple
Perforating GA- skin-coloured to erythematous papules that may have
central crust, umbilication, or pustular lesions that may exude creamy or
clear discharge
Histologically, presence of all of the following features:
a)
b)
c)
d)
necrobiosis of collagen
a predominantly histiocytic infiltrate that can be arranged in a palisading,
interstitial or sarcoidal pattern
mucin deposition which appears as basophilic stringy material between
collagen bundles
lack of epidermal change except in perforating GA which may have minor
epidermal thinning and parakeratosis as early features to complete
perforation and extrusion of the necrobiotic material
Cases must fulfill the clinical criteria 1 (a) and (b), or in case of atypical GA, 1(i)
or 1(ii) and all four histological criteria in order to be recruited into the study
C) Histopathology slide review
The pathological features of skin biopsies of GA cases were reviewed with the
generous help and supervision from four senior pathologists: Drs Lam Wing Yin (Tuen
Mun Hospital), Lee Kam Cheong (Princess Margaret Hospital), Lee King Chung (Queen
Elizabeth Hospital) and Yau Ko Chiu (Pathology Institute, Department of Heath).
5
D) Recruitment of controls
All GA cases and controls were Chinese. Controls were matched according to
age (+/-3 years), sex and clinic. They were the first three age- and sexmatched
dermatology patients in the individual clinic without the exclusion criteria listed in table
4 starting from 4th February 2003. These criteria were set to exclude dermatological
diseases associated with DM.16 If they refused to participate, family history of DM was
noted. The matching would continue until a 3:1 control to case ratio was achieved. Data
collected from controls included sex, year of birth, family history of DM and DM status
(Appendix 2). If the control did not have DM, FBS was arranged.
Table 4. Exclusion criteria for controls
1) On systemic steroid in the recent one month
or
2) Having or being suspected to have the following skin diseases:
a) Granuloma annulare
b) Necrobiosis lipoidica diabeticorum
c) Diabetic dermopathy
d) Diabetic finger pebbles
e) Scleroedema of Busckhe
f) Eruptive xanthoma
g) Reactive perforating collagenosis
h) Acanthosis nigricans
i) Neuropathic/arterial ulcers
j) Gangrene of extremities
k) Insulin induced lipoatrophy
l) Insulin induced lipohypertrophy
m) Drug eruptions secondary to
hypoglycaemic agents
E) Choice of FBS as a measure for DM
1) Reasons for using FBS
The present study employed the diagnostic criteria proposed by the American
Diabetes Association (1997). 17 It recommended that for epidemiological studies,
estimation of DM prevalence should be based on FBS. This was made in the interest of
standardization and also to facilitate fieldwork, particularly where OGTT was difficult to
perform and cost and demands on subjects’ time might be excessive. Besides its
simplicity, FBS had less variability and was more reproducible than the 2-hour value of
OGTT.17 The World Health Organization also recommended that for epidemiological
purposes, either FBS or 2-hour value after 75gm oral glucose might be used alone.18 As a
number of the GA cases were children and to facilitate the recruitment of controls, FBS
was used alone. On the basis of FBS, patients were categorized into the following groups:
Table 5. Values for diagnosis of DM and impaired fasting glucose17
6
Category
Normal
Impaired fasting glucose
Diabetes mellitus
Fasting plasma glucose
(mmmol/l)
<6.1
≥ 6.1 and < 7.0
≥ 7.0
2) Limitations of using FBS alone
Despite its convenience, the use of FBS alone would lead to a slightly lower
estimate of DM prevalence than the use of both FBS and 2-hour post value of 75gm
OGTT. In a local population-based study involving 2900 Chinese aged 25 to 74 years
using 75 gm OGTT,19 the prevalence of DM was 6.2% and 9.8% based on the American
Diabetes Association (1997) criteria17 and World Health Organization (1998) criteria18
respectively. The association of DM and GA was assessed by odds ratio. If we assumed
a similar proportion of DM was missed in both case and control by using FBS alone, the
final result, in terms of a fraction, would not be severely affected.
3) Problem of diagnostic bias
In order to reduce diagnostic bias, which meant more frequent testing of DM in
GA cases would lead to more frequent detection of DM, the first FBS ever taken in our
clinic was chosen for comparison with the controls. Moreover, age of control was
matched with age of case when the first FBS was taken. If a case had pre-existing DM,
the age of matching would be the time he had GA.
4) Reasons for not using other sensitive measures for carbohydrate
intolerance
The diagnosis of DM by FBS or by 2-hour post value of 75 gm OGTT is not
arbitrary. The prevalence of microvascular complications (retinopathy and nephropathy)
and macrovascular complication (fatal coronary heart disease) increases dramatically at
the chosen cutoff point for diagnosis of DM.17 Although steroid glucose tolerance test,
insulin secretion in response to glucose load and intravenous glucose tolerance test may
have a greater sensitivity in detecting carbohydrate intolerance, the risk of subsequent
development of DM is not known and population data are lacking in how these tests can
predict future vascular complications. Hemoglobin A1c is not recommended for
diagnosis of DM as there are different assays for glycosylated hemoglobin and further
standardization is pending.17
Impaired fasting glucose (IFG) is a recently introduced entity. Data on
cardiovascular risk of impaired glucose tolerance are mixed. Most people with impaired
glucose tolerance do not progress to DM and many may revert to normal. 20 It is
7
suggested that when impaired glucose tolerance is combined with DM to produce a
category of total glucose intolerance, an exaggerated prevalence will be obtained.20 The
same is probably true for IFG. Therefore the present study only compares prevalence of
DM in the cases and controls, without inclusion of IFG.
All cases and controls with abnormal FBS were referred to the general outpatient
clinic or specialist clinic for follow up.
E) Statistics
Mann-Whitney test was used for comparing continuous variables between groups.
Chi-square test or Fisher’s exact test was used for comparing proportions between groups.
Binomial test was employed to test significance of sex ratio. Ages of cases and controls
were compared by t-test. Mantel-Haenszel test was used for testing association of GA and
DM across age-defined strata. Logistic regression was used to find predictors for DM.
Remission data was documented by Kaplan-Meier survival curves with comparison
between curves made by Breslow test. All calculations were made by using SPSS 10.0.
Chapter 3: Results
8
A) Demographic data
Average incidence of GA from 1990 to July 2002 was 0.81 per 10,000 new cases
seen. Age of onset ranged from 2 to 81 with a median of 46 and an interquartile range of
60. As evident from figure 1, the age of onset in both male and female patients were
distributed bimodally. 38.7% of cases had onset of GA in the first 3 decades.
Figure 1. Age of onset of GA in different sex
7
6
No. of patients
5
4
3
2
Gender
1
Male
Female
0
0-9
10-19
30-39 40-49
50-59 60-69
70-79
80-89
Age range
Table 6. Age of onset of GA in relationship with gender
Age of onset
Median Interquartile Mean
range
Male
45
55.8
37.0
Female
61
62.0
38.5
Both
46
60.0
37.5
Sex
Figure 2. Sex distribution of GA cases
Gender
Male
Female
35.48%
64.52%
9
Twenty males were affected as compared with eleven females, giving a male to
female ratio of 1.82:1 (figure 2). Although there was a tendency of more male cases, the
binomial test was not significant (p=0.15). The median age of onset of GA was 45 in
male and 61 in female (table 6), which were not significantly different (Mann-Whitney
test, p=0.664).
The clinical features were documented during the interview for 16 cases (51.6%)
with active disease. For the other 15 remitted cases (48.4%), morphology of lesions was
obtained by history, clinical record and photo review. The latter group included two
cases (patient no. 15 and 26) who were interviewed by phone.
1) Classification of subtypes of GA
Cases were classified according to table 7 and the percentages of various subtypes
were shown in figure 3 (p.11). The commonest form was localized GA, accounting for
67.8%. 25.8% of cases had generalized GA. Perforating and subcutaneous subtypes
each accounted for 3.2%.
Table 7. Classification of subtypes of GA in this study
1. Localized GA
Cases of localized GA have single or multiple lesions that are confined to
one or few anatomical areas. When more than one anatomical areas are
affected, the trunk is spared.
2. Generalized GA
Generalized GA cases have extensive lesions affecting at least the trunk
and either upper or lower or both extremities.
3. Perforating GA
Perforating GA is characterized clinically by umbilicated, crusted or
pustular papules.
Histological evidence of epithelial thinning,
parakeratosis and/ or perforation with extrusion of necrobiotic material is
present.
4. Subcutaneous GA
This variant is characterized clinically by subcutaneous nodule(s).
Histologically, the necrobiotic granuloma involves the subcutaneous
layer with/without extension to the adjacent dermis.
10
Figure 3. Distribution of various GA subtypes
Subtypes
3.2%
n=1
Localized
3.2%
n=1
Generalized
Perforating
Subcutaneous
25.8%
n=8
67.8%
n=21
The age of onset in localized GA was significantly lower than that in the
generalized variant (figure 4, Mann-Whitney test, p=0.003). However, there was no
relationship between gender and subtypes of GA (Fisher’s exact test, p=1.0).
Figure 4. Age of onset of GA in localized and generalized subtypes
10
No of patients
8
6
4
subtypes
2
localized
0
generalized
0-9
10-19 30-39 40-49 50-59 60-69 70-79 80-89
Age range
11
2) Distribution
Table 8. Distribution of skin lesions in various subtypes of GA
Region
Upper Limb
Arms
Elbows
Forearms
Wrists
Hand Dorsum
Palms
Fingers
Subtotal
Lower Limb
Buttocks
Thighs
Knees
Calves/shins
Ankles
Foot Dorsum
Soles
Toes
Subtotal
Trunk
Chest
Abdomen
Back
Genitalia
Subtotal
Head and Neck
Face
Ear
Neck
Scalp
Subtotal
Localized
Number of patients affected in various subtypes
Generalized Perforating
Subcutaneous
Total
2
8
9
4
7
2
6
4
7
4
5
1
-
1
-
1
-
8
13
16
8
13
1
2
61
2
3
1
4
3
4
-
2
5
2
3
2
2
-
-
1
1
-
4
8
3
8
5
7
0
0
35
1
-
4
6
5
-
-
-
4
6
6
0
16
1
-
1
4
-
1
-
-
1
0
6
0
7
GA involved most often the upper limbs (51.3%), followed by lower limbs
(29.4%), trunk (13.4%) and head and neck region (5.9%). In the upper limbs, GA
commonly involved elbows and areas distal to them. Palms and fingers were rarely
involved. In the lower limbs, a similar pattern was observed with most lesions located
over knees and areas distal to them. Soles, toes, scalp, ears and genitalia were not
affected in this series of patients. In generalized GA, lesions could be present on the chest
wall, abdomen and back but the number of patients was too small for concluding the
12
preferential area of involvement. Only one patient had facial involvement and another
one had palmar involvement (both cases had generalized GA).
3) Morphology of individual lesion (table 9)
Predominantly annular morphology was the most common (48.4%), followed by
predominantly papular forms (22.6%) and mixed annular and papular forms (19.4%).
Epidermal change was absent in all cases except for the only one perforating GA (patient
no.9), in which the lesions occurred as tiny umbilicated papules over the hand dorsum
and neck. Subcutaneous GA (patient no. 30) presented as deep nodules over left elbow,
shin and foot dorsum. One patient described the lesions were in the form of plaque from
coalescing papules without central clearing.
Table 9. Morphology of GA lesions
Distribution of GA
Morphology
Predominantly annular
Predominantly papular
Mixed annular and papular
Perforating papules
Subcutaneous nodules
Plaque
Total
Localized
Generalized
10
7
4
1
1
23
5
2
1
8
Total
no.
15
7
6
1
1
1
31
%
48.4
22.6
19.4
3.2
3.2
3.2
100
4) Number of skin lesions
Two localized GA cases had solitary lesion (6.5%). Fifteen cases (48.4%) had
skin lesions number equal to or less than five, including one subcutaneous case. Eight
patients had generalized GA with multiple lesions, but the exact number was difficult to
ascertain. For the remaining six cases (five localized and one perforating), the total
number of lesions could not be remembered exactly but each case had more than 10
lesions.
5) Symptoms
In 24 cases, GA lesions were asymptomatic (77.4%). Pruritus was present in six
cases (19.4%), including three localized GA (one had underlying eczema), two
generalized GA and one perforating GA. One generalized GA patient, with lesions on
the first web space of hand, complained of pain on writing secondary to pressure effect.
There was no significant relationship between presence of symptoms and the subtypes of
GA (figure 5, Fisher’s exact test, p= 0.31).
13
Figure 5. Proportion of symptomatic skin lesions in various GA subtypes
No. of patients
30
20
10
SYMPTOM
Present
Absent
0
Localized
Perforating
Generalized
Subcutaneous
Subtypes of granuloma annulare
6) Precipitating factors (Table 10)
In 28 cases (90.3%), no precipitating factors could be identified. One patient with
localized GA over her upper limbs (patient no.7) claimed that the lesions occurred after
topical herbal application and that the severity of lesions increased in summer and sun
exposure. Another patient (patient no.17) with generalized GA and facial involvement
reported that the skin lesions increased with sun exposure. Patient no. 15 had GA over
upper and lower limbs and reported an increase in the number of skin lesions after attacks
of common cold and in summer.
Table 10. Reported precipitating factors for GA
Precipitating
factors
Sunburn
Drug
Common cold
Trauma
Insect bite
Weather
Pregnancy
Stress
No. of patients
2
1
1
0
0
2
0
0
%
6.5
3.2
3.2
0
0
6.5
0
0
14
C) Pathological features
A total of 32 skin biopsy specimens from 31 cases were assessed together with
senior pathologists. Necrobiosis, a predominantly histiocytic infiltration and mucin
deposition characterized the disorder and were present in all specimens. The main
histology findings were summarized in table 11. The pathological process occurred in
the dermis but the subcutaneous tissue was also involved in the only subcutaneous GA
case. The epidermis was normal except in the only perforating cases (patient no.9) in
which necrobiotic process occurred in the superficial dermis with extrusion of the
degenerated material through the epidermis. Elastophagocytosis was occasionally present
and in one case (patient no.6), there was extensive loss of elastic fibres on one side of
skin biopsy compatible with annular elastolytic granuloma, a variant of GA.
Table 11. Histology findings in GA
Pathology finding
Necrobiosis
Mucin deposition
Predominantly histiocytic infiltrate
Palisading
Interstitial
Mixed Palisading & Interstitial
Sarcoidal
Nuclear dust in necrobiotic areas
Prominent perivascular mononuclear infiltrate
DM microangiopathy
Vasculitis
%
100
100
100
58.1
25.8
16.1
0
25.8
16
0
0
D) Concurrent diseases
There were no co-existing diseases in nine cases (29.0%). The other 22 cases
suffered from one or more medical illnesses (table 12). There was no malignancy
reported. The only patient with dermatomyositis diagnosed 20 years ago (patient no.17)
did not have internal malignancy. Human immunodeficiency virus infection, although
not specifically tested in each subjects, had been checked by the medical unit for
decreased white cell count in one patient (patient no. 29) and the result was negative.
15
Table 12. Concurrent medical illnesses reported by GA cases
Medical diseases
Skin
Eczema
Pityriasis alba
Wart
Tinea pedis
Drug rash
Congenital
Down’s syndrome
Thalassaemia trait
Congenital heart disease
Metabolic/Endocrine
Gout
Thyrotoxicosis
Renal impairment
Hyperlipidaemia
Respiratory
Allergic rhinitis
Asthma
No.
3
1
2
2
1
1
1
1
1
1
1
4
1
1
Medical diseases
Cardiovascular
Hypertension
Ischaemic heart disease
Alcoholic cardiomyopathy
Varicose vein
No.
9
1
1
1
Gastrointestinal
Peptic ulcer
Intussusception
Gallstone
Hernia
Infective
Pulmonary TB
Tumour
Uterine fibroid
Benign sweat gland tumour
Others
Dermatomyositis
Osteoarthritis
Idiopathic leucopenia
1
1
2
1
2
1
1
1
1
1
E) Associated DM and previous DM screen in skin clinics
Pre-existing type 2 DM was present in five patients at the time of study. Four had
DM before onset of GA with a time interval ranging from 0.5 to 6 years (median 3 years).
One patient had DM 5 years after onset of GA. DM was diagnosed by FBS in all except
one case. The remaining one had an OGTT. No type 1 DM was identified in GA cases.
Apart from these known diabetic cases, previous screen for DM in our clinic was
present in 15 cases (57.7%, table 13). Two cases were newly diagnosed GA and hence
were not previously screened for DM. Three cases had multiple screening on follow up.
Patient no. 7 had two urine sugar tests, patient no. 8 had one spot sugar, three urine sugar
tests and patient no. 25 had two FBS. The remaining nine cases (34.6%) had never been
screened for DM in our clinics.
Table 13. Previous screen for DM and results
Methods
Urine sugar
FBS
Spot sugar
No.
8
7
5
Months after GA
(mean)
35
14.4
43.8
Results
All normal
Normal except 1 IFG
All normal
16
FBS was checked in 22 cases after interviews with the author. The remaining
nine cases consisted of five known DM cases, two cases that were interviewed by phone,
one case that had FBS two months ago when her GA was newly diagnosed and one case
that refused blood test at the time of interview but already had two FBS checked in our
clinic and the results were normal. For the two cases that were interviewed by phone,
one had normal spot sugar previously and one had never been screened for DM.
Of the 22 newly arranged FBS, one IFG and two DM cases were detected. All
three patients were referred to the general outpatient clinic for follow up.
In summary, 7 out of 30 (23.3%) GA cases had DM (one case had never been
screened for diabetes but refused to come back). Four cases developed DM before GA
from 0.5 to 6 years (median 3 years) and three DM cases were detected after GA from 2
to 5.7 years (median 5 years). One case had IFG 2.5 years after GA (figure 6). DM was
present in 15% and 50% of localized and generalized GA cases respectively. Although
the percentage of DM was higher in generalized GA cases, they had a significantly higher
age of GA onset as compared with localized GA cases. Both subcutaneous and
perforating GA cases did not have DM. The time between onset of GA and the last FBS
ranged from 1 to 206 months (mean 67.1 months).
Figure 6. Type of glucose intolerance in GA cases
DM status
DM before GA
DM after GA
IFG
12.5%
37.5%
50.0%
17
F) Case-control study of association with DM
The case-control study included only those cases with three matched controls
found. Twenty-two cases were eligible, including thirteen localized, seven generalized,
one perforating and one subcutaneous GA. The nine non-eligible cases consisted of two
cases who were interviewed by phone and another seven cases without enough matched
controls. The latter seven non-eligible cases did not have DM at first FBS screen. 9 out
of 66 controls had family history of DM. This was not significantly different from the
proportion of those who refused to participate, in which 5 out of 59 had family history of
DM (chi-square test, p=0.36).
Eight controls had pre-existing DM (all diagnosed by FBS except two by spot
sugar and one by OGTT). Four new DM and two new IFG cases were found on
screening. Both groups were referred for further management. The mean age of cases
and controls were 52.2 (standard deviation (SD) 22.8) and 52.1 (SD 22.2) respectively,
which were not significantly different (t test, p=0.99).
The distribution of DM in cases and controls were shown in table 14. They were
stratified into 3 age ranges. The ranges were chosen to reflect the prevalence of DM in a
local population based survey,21 which showed an age related increase in DM prevalence.
Table 14. Distribution of DM in cases (at first FBS screen) and controls
Age range
0-34
35-54
55-86
Group
Case
Control
Case
Control
Case
Control
DM
Absent Present
5
0
17
0
4
1
11
2
7
5
26
10
% DM
Total
0
0
20
15.3
41.7
27.8
5
17
5
13
12
36
Although the percentage of DM in cases seemed to be greater than that of controls,
the difference was not statistically significant (Mantel-Haenszel test, p=0.56). The
Mantel-Haenszel odds ratio was 1.75 with a 95% confidence interval (CI) of 0.5 to 5.8.
Logistic regression did not demonstrate presence of GA (p=0.36) and family history of
DM (p=0.23) as significant factors in predicting DM. Subgroup comparison of localized
and generalized GA with their respective controls did not reveal any significant
association with DM (Mantel-Haenszel test, p=0.49 for localized GA, p=0.83 for
generalized GA). There was no difference in family history of DM between cases and
controls (Fisher’s exact test, p=1.0).
Two previous studies based on DM status (rather than sensitive tests for
carbohydrate intolerance) gave a relative risk of 11.4 in one10 and an odds ratio of 2.6 in
another.13 If we assumed a relative risk of 5, 18.2% DM prevalence in controls (12/66),
case-control ratio of 1:3 and α= 0.05, the power of study was 88%.
18
G) Family history of GA and DM
Family history of GA was absent in all cases. Three cases had positive family
history of DM, in which two had DM themselves and one was free from DM at the time
of the interview.
H) Duration of follow up
The cases were followed up for a range of 14 to 206 months after the onset of GA
with a mean and median of 67.5 and 68 months respectively. Over 80% of cases were
followed up for more than 24 months.
I) Treatment (Tables 15, 16)
Topical steroid (mostly of moderately potent or potent class according to the
classification in Martindale22) was the most often employed treatment in this series (28
cases, 90.3%). In 16 cases, complete remission (CR) had been achieved with topical
steroid alone. No treatment had been offered to three localized GA patients (patient no.
10, 21 and 24) as CR occurred within one month after skin biopsy. Six patients (patient
no. 4, 9, 12, 16, 26, 27) had CR within two months after skin biopsy together with the use
of topical steroids. Their GA lesions were disseminated, making it difficult to ascribe CR
to biopsy alone. In patient no. 25, CR had been achieved by a combination of topical
steroid and intralesional steroid.
Oral dapsone (50 mg/day) was used for five months in a paediatric patient
(patient no. 30) with subcutaneous GA but no clinical response was noted. She was then
treated with topical steroid and intralesional steroid for 5 courses and had first CR after
70 months. Patient no. 6 with generalized GA had persistent disease after five months of
moderately potent topical steroid. Oral isotretinoin 0.5mg/kg/day was then given,
resulting in CR after six months. Side effect of isotretinoin in this patient included mild
cheilitis. The liver function and lipid profile remained normal during the treatment.
Table 15. Response to treatment in GA cases
Treatment modality
No. of cases
CR by skin biopsy alone
3
CR by topical steroid alone
16
CR by topical and intralesional steroid
1
CR by isotretinoin
1
PD with dapsone, CR by topical and intralesional steroid
1
PD after topical steroid
7
Newly diagnosed, just started topical steroid
2
Total
31
Abbreviations: CR: complete remission; PD: persistent disease
19
Table 16. Summary of treatment modalities used in GA cases
1
2
3
LGA
LGA
LGA
Potent TS
Moderately potent TS
Duration of
treatment
(months)
24
Just started
2
4
5
LGA
LGA
Potent TS
1
7
7
LGA
8
10
11
15
LGA
LGA
LGA
LGA
46
4
12
20
16
18
19
LGA
LGA
LGA
Potent TS
Potent TS
Potent TS
Nil
Potent TS
Potent TS
Mild TS
Potent TS
Potent TS
Potent TS
21
23
LGA
LGA
24
26
27
Patient
no.
GA
Subtype
Treatment
Outcome
CR
Newly diagnosed
Defaulted, resolved after 24
months
CR after 1 month
CR. Disease free for 10 months
39
3
2
2
Just started
14
PD, defaulted FU 40 months
PD, defaulted FU 11 months
PD
CR
Subsided 1 month after biopsy
PD
Defaulted FU, CR after 14
months
CR
Newly diagnosed
CR, disease free for 9 months
Nil
Very potent TS
Potent TS
6
20
Subside 1 month after biopsy
CR after TS, disease free for 28
months
LGA
LGA
LGA
Nil
Potent TS
2
1
1
28
LGA
Potent TS
13
29
31
6
LGA
LGA
GGA
4
4
5
6
12
13
GGA
GGA
Potent TS
Potent TS
Oral isotretinoin (0.5
mg/kg/d)
Potent TS
14
GGA
12
17
20
GGA
GGA
Potent TS
Potent TS
7
4
22
25
GGA
GGA
9
30
PGA
SCGA
Potent TS
Moderately
potent TS,
steroid x 1
Potent TS
Dapsone
Moderately
potent TS,
steroid x 5
potent and
intralesional
potent and
intralesional
2
5
14
25
2
5
70
CR 1 month after skin biopsy
CR
CR after two months. Then had
multiple
recurrences
and
remission, patient could not
remember the details. Now
disease is still present
PD, defaulted FU. Disease is still
present now
CR
Defaulted FU, PD
PD with TS
CR after 6 months of isotretinoin
CR
Defaulted FU, PD all along,
given TS on reassessment
Default FU, CR after 43 months.
In remission for 4 months
CR
PD, but defaulted FU, given TS
again on reassessment
PD
CR, disease free for 6 months
CR
No response to dapsone then CR
after TS and intralesional steroid,
disease free for 12 months
Recurrence
Nil
Nil
Nil
Nil
Recurred for 10 months,
put on potent TS again
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Recurred 1 month recently,
just restarted potent TS
Nil
Recurred in recent 2
months, just
restarted
potent TS
Nil
Nil
Yes, but defaulted FU, just
started potent TS again
Nil
Nil
Nil
Nil
Nil
Nil
months,
just
started
moderately potent TS
Nil
Nil
Nil
months, put on potent TS
Nil
Yes, relapse treated with
potent TS, intralesional
steroid x 1. CR again 15
months later
Abbreviations: LGA, GGA, PGA, SCGA: localized, generalized, perforating and subcutaneous subtypes of
GA respectively. TS: topical steroid; FU: follow up; CR: complete remission; PD: persistent disease
20
J) Prognosis
CR had been achieved in 21 out of 31 (67.7%) GA cases. The median time for
first CR was 43 months (CI 24.4-61.6 months) for all cases as a whole (figure 7). The
median time for CR was 54 months (CI 11.3-96.7 months) for generalized GA and 39
months (CI 13.6-64.4 months) for localized GA, but the difference was not statistically
significant (figure 8, Breslow test, p=0.13). Likewise, despite cases with DM seemed to
last longer than non-diabetic cases (figure 9), it was not significant (Breslow test p=0.37).
No significant difference was observed in chances of first CR between age of onset < 20
and age of onset ≥ 20 groups (figure 10, Breslow test, p=0.73) and between male and
female (figure 11, Breslow test, p=0.89).
Figure 7. Proportion of first CR of all GA cases with time
1.2
Proportion in first remission
1.0
.8
.6
.4
.2
0.0
Survival Function
Censored
-.2
0
20
40
60
80
100
120
140
Duration of disease (months)
Figure 8. Proportion of first CR in localized and generalized GA
1.2
1.0
.8
.6
Subtypes of GA
.4
General
General-censored
.2
Local
Local-censored
0.0
0
20
40
60
80
100
120
140
21
Figure 9. Effect of diabetic status on first CR of GA
1.2
1.0
.8
.6
.4
Diabetes mellitus
.2
present
present-censored
0.0
absent
-.2
absent-censored
0
20
40
60
80
100
120
140
Figure 10. Effect of age of onset on first CR of GA cases
1.2
Proportion of first remission
1.0
.8
.6
.4
Age group
.2
>20
>20-censored
0.0
<20
-.2
<20-censored
0
20
40
60
80
100
120
140
22
Figure 11. Effect of sex on first CR of GA cases
1.2
1.0
.8
.6
.4
Gender
Female
.2
Female-censored
0.0
Male
-.2
Male-censored
0
20
40
60
80
100
120
140
7 cases out the 21 cases (33.3%) recurred after the first CR. These comprised
four localized, two generalized and one subcutaneous GA. The median time for relapse
was 9.5 months. Only one patient in the relapsed group had DM. Patient no. 27 with
localized GA had multiple recurrences but she forgot the exact details and duration. The
range of follow up time after relapse was 1 to 15 months. Six relapsed cases had
persistent disease at the time of interview. Only the subcutaneous case had second
remission after treatment with topical and intralesional steroid.
23
Chapter 4: Discussion
A) Epidemiology
An incidence of 0.12% new patients was reported in one hospital 23 and an
incidence of 0.1% to 0.4% new patients was reported in one review article.24 An average
incidence of 0.0081% new cases is calculated in the present study, which is an
underestimation because not all clinical records are retrievable. In contrast with the
literature (table 17), this study seems to have more male cases, but this is not statistically
significant.
Table 17. Age and sex distribution of GA in the literature.
Source
Localized GA (%)
Studer EM, et al3
Tan HH, et al4
Muhlemann MF, et al10
Wells RS, et al23
Dabski K, et al25
This study
75
82.9
93
95
0
67.8
Age of onset
(% in first 3
decades)
62
66
86
71
11
38.7
Male to
female
1: 2.8
1: 1.56
1: 1.25
1: 2.2
1: 2.2
1.82:1
Onset of GA was mostly in the first three decades in the literature,3,4,10,23
contrasting with only 38.7% of cases in this study. The difference can be attributed to
the variation in proportion of GA subtypes. In a series of patients with generalized GA,25
only 11% had their disease onset in the first three decades. In this study as well as in
others,26 the age of onset of GA is higher in the generalized than the localized variant.
This study has 67.8% of localized GA, as compared with 75-95% in other series,3,4,10,23
and this may be the cause of the later onset of disease in our series. A bimodal age of
onset described for generalized GA24 is not seen in the present study in which a single
peak of onset in the sixth decades is present.
Subcutaneous GA occurs most often in children and young adults,2,24 whereas in
perforating GA, a later age of onset has been described. 27 In the present study,
perforating and subcutaneous cases have age of onset at four year and three year
respectively.
Familial GA has been reported in the literature,23,28 but is absent in our series.
Although there is no satisfactory explanation, a genetic component via human leucocyte
antigen (HLA) association has been reported (see Etiology).
24
1) Morphology of skin lesions
Dabski K et al has divided the clinical appearance of generalized GA into two
forms: the predominantly annular form (67%) and the predominantly nonannular form
(33%).25 In the present study, the predominantly annular form accounts for 48.4%,
whereas the predominant papular form accounts for 22.6%. Some patients have a mixed
papular and annular pattern and this accounts for another 21.2%. One patient has
coalescing papules forming a plaque without central clearing. This has been classified as
nonannular form in Dabski K et al series but the term “plaque form” may be more
descriptive for the clinical appearance.
In perforating GA, umbilicated and crusted lesions are the commonest (> 70%),
while pustular lesions occur in 26% of cases.27 Lesions may heal with atrophic hyper- or
hypopigmented scars. Clinically, perforating GA is a disseminated disease with multiple
papules (only 9% has single lesion) and can be classified, according to their distribution,
into localized and generalized variant. Only one case of perforating GA is present in this
study, presenting with umbilicated papules over neck and hand dorsa.
Subcutaneous GA usually presents as painless skin- coloured deep dermal or
subcutaneous nodules.2,24 It is also called pseudorheumatoid nodule or benign rheumatoid
nodule but no relationship with rheumatoid arthritis is observed. Superficial papular
lesions are present in about one fourth of cases.24 One subcutaneous GA is present in this
study with painless skin –coloured papules distributed over the extremities.
2) Classification
The definitions of various subtypes of GA are listed in table 7. Localized GA has
solitary or a limited number of skin lesions that may affect one or several anatomical
areas5 and papules may be arranged in an annular or arciform pattern.29 Generalized GA,
according to Dabski K et al,25 has widely distributed lesions of a large number affecting
the trunk and at least one extremity. Tan HH et al has used the term disseminated GA to
describe extensive lesions confined to one anatomical site or to the extremities only.4
But this is the only case series, according to the author’s knowledge, classifying
disseminated GA as distinct from localized and generalized type. The exact number of
skin lesions that defined ‘extensive’ was not mentioned in their article and most other
case series in the literature did not include a disseminated subtype.3,9,10 Therefore this
subtype is not employed in the present study. Subcutaneous2,24 and perforating27
subtypes differ from typical GA by the level of skin involvement: subcutaneous layer in
the former and superficial dermis as well as epidermal involvement in the later. This
pathological difference translates into difference in clinical characteristics of skin lesion
as outlined above.
25
The percentages of each GA subtypes in this study and in various series are
similar (table 18). Localized GA is the most frequent, followed by generalized subtype.
Perforating and subcutaneous forms account for less than 10%. This study, as compared
with other series, has a lower proportion of localized GA. Generalized GA accounts for
25.8%, which lies in the upper range of percentage as compared with the literature (525%, see table 18). Perforating and subcutaneous GA both occur in 3.2% in this study,
which is comparable to 1.2% and 4.8% respectively reported by Studer EM et al.3
Table 18. Subtypes of GA in various case series
Source
Studer EM, et al3
Tan HH, et al4
Haim S, et al9
Muhlemann MF, et al10
Wells RS, et al23
Dabski K, et al25
This study
Year LGA %
1996
75
2000
82.9
1973
75
1984
93
1963
95
1989
0
2002
67.8
GGA %
19
9.8
25
7
5
100
25.8
PGA %
1.2
2.4
0
0
0
0
3.2
SCGA %
4.8
0
0
0
0
0
3.2
DGA %
4.9
-
Abbreviations of GA subtypes: LGA: localized subtype; GGA: generalized subtype; PGA: perforating
subtype; SCGA: subcutaneous subtype; DGA: only coined in one study, represent disseminated subtype
3) Other variants of GA reported in the literature
i) Papular umbilicated GA
It was described in four paediatric cases.30 All occurred over dorsa of hands as
firm umbilicated papules. Histologically, necrobiotic granulomas were present beneath
areas of epidermal thinning and parakeratosis. There was no perforation. Similar
epidermal changes were reported in early perforating GA.27 The authors acknowledged
that there was a disease spectrum from papular to perforating subtype and it appeared to
be a variant of perforating GA.30
ii) Patch GA
It presented as asymptomatic erythematous to brown patches, without papules,
scaling or induration.31 Histologically, an interstitial histiocytic pattern was observed. 4
It was
out of 6 patients in the original series were on long-term medications.31
questioned as a genuine variant of GA, but rather represented an example of interstitial
granulomatous form of drug reaction.29
iii) Follicular pustulous GA
It presented as recurrent eruptions of generalized pustules over trunk and palms.32
Histologically, palisading necrobiotic granulomas occurred in a perifollicular distribution.
Intense necrosis of follicular walls and neutrophilic infiltration of upper portions of
26
follicles were noted. Transepidermal perforation of the necrobiotic material was seen in
palmar lesions. The adnexal affinity of the granulomatous process and superficial
aggregation of neutrophils in follicular opening suggested a pattern different from
ordinary GA. But in the presence of perforation, the authors could not exclude the
possibility that it belonged to perforating GA.32
iv) Linear GA
It presented as a linear subcutaneous band over the lateral chest wall.33 This entity
is now regarded as a variant of interstitial granulomatous dermatitis with arthritis that can
present as linear cords (‘rope sign’), papules or plaques.29 The disease is associated with
rheumatoid arthritis. Histologically, as compared with GA, the histiocytes are usually
scattered interstitially, especially marked in the lower dermis. Moreover, eosinophils and
neutrophils are more noticeable and mucin deposition is less in this variant.29 In some
areas, a few degenerated collagen bundles are enveloped by neutrophils, eosinophils and
histiocytes, forming Churg Strauss granuloma.
v) Actinic granuloma
In 1975, O’Brien described this entity, characterized clinically by annular plaque
on sun-exposed areas; histologically by absence of elastic fibres in the post reactive
central zone and granulomatous inflammation with giant cells at the edge of lesion.34 The
uniqueness of the condition and the concept of granulomatous response to solar elastosis
were controversial. Ackerman AB considered it as a variant of GA that occurred on sun
exposed skin.35 Annular elastolytic giant cell granuloma was proposed by Hanke CW et
al in 1979 to include the entity actinic granuloma. 36 They agreed that the mere
occurrence of solar elastosis and adjacent granulomatous inflammation did not imply a
cause-effect relationship. Further argument will be continued, unless the etiology of each
condition is found.
4) Distribution of skin lesions
The distribution of skin lesions is similar to that reported in the literature.3,4 Distal
extremities beyond the elbow or knee region are commonly affected. In perforating GA,
the extremities especially the hands and arms are commonly involved.27 Subcutaneous
GA commonly involves not only distal extremities but also scalp and forehead.37
5) Symptoms (Table 19)
Skin lesions of GA are basically asymptomatic. Pruritus and pain may be present
in some cases. A study on generalized GA showed a higher proportion of symptomatic
cases.25 Our study also shows a similar trend. Symptoms are present in 37.5% and
14.3% of generalized and localized GA cases respectively, but the difference is not
significant (Fisher’s exact test, p=0.31).
27
Table 19. Proportion of symptomatic GA cases in the literature
Source
Tan HH, et al4
Dabski K, et al25
This study
GGA %
9.8
100
25.8
Asymptomatic% Symptomatic%
85.4
14.6
66
34
78.8
21.2
Abbreviation: GGA: generalized granuloma annulare
Most patients cannot identify any precipitating factors. Reported precipitating
factors in the literature were sunburn, trauma, insect bite, drug, flu-like illness, reaction to
intravenous contrast medium, phlebitis, sepsis after surgery and stress.3,4,25 The present
study also identifies some precipitating factors but cannot prove a cause-effect
relationship or the mechanism involved in the exacerbation.
C) Pathology
1) Necrobiosis and mucin deposition
GA is characterized by necrobiosis, granulomatous inflammation and mucin
deposition. Necrobiosis refers to non-specific connective tissue alteration. It appears as
blurring and loss of definition of collagen bundles, separation of fibres, a decrease in
connective tissue nuclei and an alteration in staining by routine histological stains, often
with increased basophilia or eosinophilia.29, 38 Necrobiosis is usually minimal in
interstitial GA and careful scrutiny is required. Mucin appears as bluish stringy and
granular material on hematoxylin and eosin stain. It can be enhanced by alcian blue or
colloidal iron stain.
2) Inflammatory cells
The inflammatory cellular component consists of predominantly histiocytes (some
of which are epitheloid or multinucleated), lymphocytes (majority of which are T helper
cells) and occasionally eosinophils.29,34 Table 20 highlights the difference in histiocytic
pattern between this study and the literature.
Table 20. Pattern of histiocytic infiltration in GA
Histiocytic infiltration
Palisading
Interstitial
Mixed palisading & interstitial
Sarcoidal
Umbert P et al38
%
71
26
3
This study
%
58.1
25.8
16.1
-
28
Moreover, Dabski K et al found that generalized GA, in comparison with
localized subtype, had less apparent collagen abnormality and fewer cases with
palisading pattern.39 Our study does not demonstrate such differences. The reason for
difference of histology pattern between the present study and that of the literature may be
related to the focal nature of the disease. Some skin biopsies show a mixture of
histiocytic infiltration patterns and there are variable histology patterns depending on the
sections.
3) Elastic fibre changes
Elastic fibre reduction was present in 20-30% of GA cases in one series.39 In our
study, one case has almost complete loss of elastic fibres on one side of biopsy specimen
compatible with annular elastolytic giant cell granuloma. Degenerated elastic fibres may
initiate a granulomatous response or it may be lost secondary to the activity of elastase or
phagocytosis of inflammatory cells. Clinically atrophic patches with histological features
of mid-dermal elastolysis40 or anetoderma41 have been reported after resolution of GA.
4) Vascular changes
A perivascular lymphohistiocytic infiltrate is frequently found24,35,38 and in 16 %
of cases in this study, this is especially prominent. Diabetic microangiopathy has been
found in lesions of GA especially in the generalized subtype26 but this feature is absent in
our study even in cases with known DM.
Vasculitic changes and positive
immunofluorescence for C3 and IgM (in up to 50% of cases) had been reported and the
author concluded vasculitis was possibly involved in the pathogenesis of GA.42 However,
the results were not reproducible, immunofluorescence for IgM and C3 was positive on
blood vessels in only 9% of cases in another study.43
It has been suggested that the occurrence of nuclear dust in the necrobiotic area
indicates vasculitis is an early event and may occur at the center of palisading granuloma,
thus it may be missed by random section or biopsy done late in the disease course.35,44
To test the hypothesis, skin biopsy with immunofluorescence study of early (< 1 week)
localized GA and normal skin adjacent to spreading GA was performed.44 The direct
immunofluorescence was negative in all specimens and the result did not support the
presence of vasculitis.
In the present study, nuclear dust is present in 25.8% of specimen and this may
represent apoptotic histiocytes, fibroblasts or other cell types (personal communication
with Dr. KC Lee, PMH). Vasculitic changes are absent in all specimens but direct
immunofluorescence has not been carried out.
5) Subcutaneous and perforating GA
Subcutaneous GA shows essentially the same histology located at subcutaneous
tissue and deep dermis. Eosinophils are more commonly seen29 and interstitial pattern is
not observed in the septa.35 Perforating GA has necrobiotic granuloma located at
29
superficial dermis.27 Epidermal changes range from epithelial thinning, parakeratosis and
perforation in the epidermis or follicular wall. There is controversy on whether the
perforation is simply epidermal destruction by the inflammation process or the material is
transepithelially eliminated.
D) Etiology
1) Genetic Factors
GA has been reported in families and twins.24 Friedman –Birnbaum R et al
reported increase HLA-A31 and B35 in Ashkenazi Jews of Eastern European origin in
generalized but not in localized GA.45 The authors have made comparison with other
studies as well. HLA-A29 is increased in patients with localized GA from the Belfast
area. Increase HLA- B8 is found in Danish patients with localized GA. Because HLAB8 is associated with DM, this is the only study that points to a possible genetic
association of DM with GA.
2) Physical Factors
Trauma has been reported in 25% of children in subcutaneous GA.24 The
appearance of papular GA on exposed areas has prompted some authors to consider
insect bite as a cause.24 Sun exposure and photochemotherapy have been reported as
provocating factors.1
3) Diseases associations
The diseases reported in association with GA are summarized in table 21. These
are mostly based on case report or small case series and there are neither direct evidence
nor a proven mechanism of the diseases association. In the present study, apart from DM,
only two cases have past history of pulmonary tuberculosis. The other diseases listed in
table 21 are not present.
Table 21. Reported diseases associated with GA
Factors
Infections
Varicella Zoster
virus
Epstein Barr
Virus (EBV)
Hepatitis C
virus (HCV)
Article
Case report
Case report48
Case report50
Postulated mechanisms/ Remarks
Viral DNA detected by PCR in GA lesions,46 but
this was not reproducible by others47
Original reported case was probably not GA.49
Subsequent study using PCR and ISH failed to
demonstrated EBV in GA49
Liver epitheloid granulomas are present in 10%
of HCV related cirrhosis, indicating the ability of
virus in inducing granulomatous response
30
Table 21 (continued). Reported diseases associated with GA
Factors
Human
Immunodeficien
cy virus (HIV)
Systemic
diseases
Article
Case
series49,51
Mycobacteria
Borrelia
burgdorferi
Autoimmune
thyroiditis
Case report52
Case series53
Sarcoidosis
Case series55
DM
Sweet’s
syndrome
Case report56
Morphoea
Case report57
Temporal
arteritis
Case report58
Malignancy
Case series59
Case control
study54
Postulated mechanisms/ Remarks
Characterized by a predominant CD8 positive
lymphocytic infiltrate.
Atypical presentations such as perforating GA,
macular erythema may be present.
One case series showed a higher generalized GA
presentation (59%).49
Skin biopsy is advisable in HIV positive cases to
exclude other differential diagnosis.
Nil
Urine PCR for Borrelia positive in 60% of GA
cases
Associated with localized GA in female, a
common immunogenetic predisposition or
immunopathogenetic mechanism proposed
Both are characterized by granulomatous
inflammation, delayed hypersensitivity to
obscure antigens and macrophage migratory
inhibitory factor is found in both. Presence of
increased dermal mucin or eosinophils in GA
may be helpful distinguishing features.
See table 1 and p.32
Generalized GA and Sweet’s syndrome, both can
be associated with malignancy, occur in a patient
with breast carcinoma.
Autoimmune etiology is proposed; blood vessel
damage and upregulation of collagen synthesis
are present in both diseases.
Both are characterized by granulomatous
inflammation, giant cells and loss of elastic
fibres and sunlight has been proposed as a
causative factor in both.
Association with Hodgkin’s disease, nonHodgkin lymphoma, Sezary syndrome,
Lennert’s lymphoma, leukaemia, carcinoma of
lung, cervix, colon, breast, prostate, stomach,
thyroid and seminoma have been described. GA
occurs 14 years before to 27 years after
malignancy. No definite relationship is
established. Occurrence of painful lesions or
palmoplantar and facial lesions is said to be
more frequent in this group.
Abbreviations: PCR: polymerase chain reaction; ISH: in-situ hybridization
31
4) Association with DM
Despite a substantial number of literatures linking GA with DM, the present study
fails to detect such an association. Case-control studies like the present one only imply
association with or without causation. In the absence of experimentation, several criteria
have been advocated for assessing causality:60
i) Temporal sequence
A casual association requires causative factors precede results. In the present
study, 3 patients had diabetes mellitus after GA, such occurrence has also been
reported in other series.10
ii) Consistency and strength of association
Summary of some of the reported studies has been given in table 1. No consistent
association has been demonstrated. Even in studies showing association, the
strength of association is not consistently high (a relative risk of 11.4 in one10 and
an odds ratio of 2.6 in another13).
iii) Dose-response relationship
There has not been any study on the relationship between severity of GA and the
degree of hyperglycaemia. Although reports on resolution of GA after
chlorpropamide are present,7 definite conclusions cannot be drawn as they are
uncontrolled.
iv) Biological plausibility
The mechanism of DM causing GA is rarely studied. This may be due to the fact
that the pathogenesis of GA itself is not known. The possible mechanisms are:
a) Immunogenetic linkage
HLA B8, B15, DR3 and DR4 are associated with DM, whereas HLA DR2 is
‘protective’ against development of DM. The only series showing possible
association of GA and DM is increased frequency of HLA-B8 in Danish patients
with localized GA, which is not reproducible in other populations.
b) Diabetic microangiopathy
Diabetic microangiopathy has been found in high frequency in generalized GA
but not in localized type.26 Palisading granulomas are suggested to be a
consequence of tiny infarcts that cause collagen degeneration, to which
histiocytes respond in a radial array. However in a large study on histology of
GA38 and in 100 patients with generalized GA,39 diabetic microangiopathy is
absent. There is no diabetic microangiopathy identified in the present study.
c) Factor VIII related antigen
High levels of factor VIII related antigen are linked with microvascular disease in
DM. Factor VIII related antigen is raised in necrobiosis lipoidica and widespread
GA, even in patients without DM, but not in localized GA.61 Its enhancement of
32
platelet adhesiveness and aggregation may contribute to vascular injury. But
evidence is lacking on whether the raised factor is the result of the disease process
rather than the cause.61
In summary, there is no strong evidence of DM as a causative factor in GA.
5) Drugs
Vaccinations including BCG, hepatitis B and antitetanus vaccines have been
reported to induce GA. Vaccination may have initiated GA through traumatic or
immunological mechanisms.62
Drug induced GA has been reported and drugs such as gold, allopurinol,
calcitonin and diclofenac are incriminated.63 An entity, interstitial granulomatous drug
reaction, has been described in 1998.29 The reported culprits included calcium channel
blockers, beta-blockers, angiotensin converting enzyme inhibitors, lipid-lowering drugs,
antihistamines, anticonvulsants and antidepressants.
Clinically, skin lesions are
erythematous to violaceous plaques, often with annular configuration, affecting medial
arms, thighs, intertriginous areas and trunk. The histology mimics interstitial GA. They
are differentiated from GA by the absence of significant necrobiosis, predominance of
lymphocytes over histiocytes, presence of eosinophils and interface changes with mildly
atypical lymphocytes and focal epidermotropism.
E) Pathogenesis
1) Cell mediated immunity
Infiltrates of GA are characteristics of cell-mediated immunity, with histiocytes
and variable lymphocytes (T helper> T-suppressor). 64 Serum macrophage inhibitor
factor activity has been detected in patients with GA.65 Its detection can be used as a
correlate for cellular immunity status. Peripheral blood leukocyte function, including
polymorphonuclear and monocyte chemotaxis, phagocytosis, oxidative burst and
blastogenic response of lymphocytes have been found to be normal in one study in
vitro.66 But functional abnormality of leukocytes in vivo cannot be excluded.
The nature of the infiltrating T cells has been studied. 67 There is a selective
infiltration of a limited number of skin-specific T cells clones together with a large
number of non-specific T cells. This situation reflects a T cell infiltration model
previously proposed for autoimmune diseases. A predominant interleukin-2 transcription
is also found which may provide signal for ongoing inflammation and cellular
recruitment.
Besides interleukin-2, interferon-gamma (IFN-γ) is expressed by a large number
of lymphocytes, compatible with a Th1 cytokine pattern. Other cytokines involved are
the tumour necrosis factor- alpha, interleukin-1 beta, interleukin-6 and matrix
metalloproteinases. They all contributed to the inflammatory response in GA.68
33
It is proposed that sensitized lymphocytes in dermis release cytokines, resulting in
accumulation and activation of macrophages and hence release of various enzymes that
may contribute to connective tissue degeneration.64 Potential antigens may include
altered collagen, altered elastic fibres, bacterial/viral agents and antigens in saliva of
arthropods.2
2) Immune complex vasculitis
Detailed discussion has been given in pathology section. In general, this is not a
favoured mechanism as subsequent studies failed to show the presence of vasculitis in as
a high proportion as in the series of Dahl MV et al.42 Furthermore, in inflammatory
dermatosis, complement and immunoglobulin may sometimes be deposited in vessels
without true vasculitis.24
3) Collagen abnormalities
The synthesis of type I and III collagens are increased in GA and areas of
collagen degeneration are surrounded by active collagen synthesizing fibroblasts. The
number of proα1(I) collagen mRNA molecules is increased four to five fold as compared
with nonlesional skin. This represents a reparative phenomenon in GA.69
The expression of various enzymes affecting collagen turnover in GA have been
studied. 70 Interstitial collagenase is expressed in the peripheral rim of histiocytes in
palisading granuloma of less than 1 year old. This spatial relationship suggests a
centrifugal process whereby collagenolysis occurs, starting from the centre and
expanding outward with active degeneration corresponding to the site of expression of
interstitial collagenase, that is, rim of granuloma. This is not observed in interstitial
granuloma annulare in which collagenase expression is consistently negative. Tissue
inhibitor of metalloproteinases (TIMP)-1 regulates activitiy of collagenase. TIMP-1 is
expressed in similar areas as interstitial collagenase and is present in young and old GA
lesions. This inhibitor expression may be a homeostatic response to collagenase, to limit
the damage and to allow for repair. 92-kDa gelatinase is expressed by eosinophils with
no relationship to the granulomas, casting doubt on its role in elastolysis in the
necrobiotic centre.
F) Differential diagnosis
•
•
•
•
•
•
•
The differential diagnosis includes:
Tinea corporis
Subacute cutaneous lupus erythematosus
Erythema marginatum
Annular lichen planus
Erythema multiforme
Erythema annulare centrifugum
Erythema migrans
34
•
•
•
•
•
•
Leprosy
Nodular tertiary syphilis
Sarcoidosis
In non-annular forms, cutaneous amyloidosis, scleromyxedema, papular
mucinosis and erythema elevatum diutinum
In subcutaneous form, rheumatoid nodules and other deep cutaneous tmours
In perforating form, acquired reactive perforating dermatoses, perforating
folliculitis, elastosis perforans serpiginosa, molluscum contagiosum and
papulonecrotic tuberculid
G) Treatment
GA is rare and there is no large-scale double blind randomized controlled trial on
treatment. Together with the fact that spontaneous resolution may occur, it is difficult to
interpret the reported efficacy of various treatments. Some authors stated that treatment
of GA was generally unnecessary, owing to their spontaneous resolution.24 But the
persistence and disfigurement of generalized GA has prompted treatment in other cases.71
The choice of therapy depends on GA subtypes, side effects and patient preference.
Table 22. Therapeutics for GA (modified from reference72)
Localized GA
Generalized GA
First line
First line
Cryotherapy
Photo(chemo) therapy
Intralesional and topical steroid
Systemic isotretinoin
Dapsone
Second line
Second line
Intralesional interferon
Cyclosporine
Surgery
Systemic steroid
Topical imiquimod
Chlorambucil
Cream PUVA
Antimalarials
Potassium iodide
Pentoxifylline
Niacinamide
Vitamin E and 5-lipoxygenase
inhibitor
Fumaric acid esters
1) Resolution after skin biopsy
Reports of resolution of GA after skin biopsy are present.31, 73 As they are
uncontrolled, it is difficult to eliminate the element of natural regression. Three cases of
localized GA in the present series resolved within one month after skin biopsy. The
possible mechanism of biopsy-induced resolution is by initiating an orderly process of
wound healing involving interactions of epidermal, inflammatory, fibroblastic and other
cell types through cytokines.73 However, it is not clear why only some GA cases resolve
after biopsy.
35
2) Steroids
91.2% and 5.9% of localized GA in one series (34 cases in total) were treated
with topical and intralesional steroids respectively. 67.6% of patients had remission after
one year of follow up.4 In another study,74 intralesional triamcinolone acetonide produced
complete clearance in 68% (out of 25 cases) as compared with 44% (out of 27 cases)
injected with normal saline after a mean treatment sessions of 2.4 at 6 to 8 weekly
intervals. Lesions relapsed in about 42 to 47% in both groups. Transient atrophy at
steroid injection sites was common.
The response to steroid therapy is consistent with the immune pathogenesis of GA.
In the present study topical steroid is the most common treatment resulting in remission
either alone (16 cases) or together with intralesional steroid (1 case). Seven cases have
persistent disease with topical steroid.
3) Cryotherapy
One freeze thaw cycle of 10-60s per session was used in 31 localized GA patients
using either nitrous oxide or liquid nitrogen and treatment was repeated every 20 to 30
days.75 80.6% of patients had remission after a single treatment. Relapse occurred in 9%
of patients who were followed up for more than two years. Atrophic scarring occurred in
21.1% of patients treated with liquid nitrogen but was absent in those treated by nitrous
oxide.
4) Intralesional interferon
Three localized GA patients were given intralesional recombinant human
interferon gamma (IFN-γ) 2.5x 105 IU/lesion on seven consecutive days and thereafter
three times per week for two more weeks.76 As a control, other lesions were treated with
normal saline or steroid under occlusion. After three weeks, all IFN-γ treated lesions
improved but controls did not. Possible mechanism involves inhibition of delayed
hypersensitivity reaction by IFN-γ.
5) Surgery
Scarification, using injection needle drawn across lesions to produce capillary
bleeding, was used successfully in two cases.77 In addition, razor blade subsection with
healing by secondary intention was used in a patient with ‘nodular’ GA over fingers.78
6) Topical imiquimod
Topical imiquimod nightly for six weeks resulted in resolution of GA in a case
report.79 Other lesions that were not treated remained the same.
7) Cream psoralen plus ultraviolet A (cream PUVA)
Increased sensitivity to UVA was accomplished by topical psoralen cream
(0.001% 8-methoxypsoralen). In this way, only selected body parts were included in the
36
therapy (with possible decrease in side effects). 80 Cream PUVA was used in five
localized GA patients four times a week starting at 30% minimal phototoxic dose, with
increments at 30% of last UVA dose and maximum increment at 0.5J/cm2. The treatment
was continued up to two sessions after resolution. Four showed complete response while
one showed significant improvement after a mean of 26 treatments (mean UVA dose
55.9J/cm2). No recurrence was noted after four months of follow up.
8) Photo(chemo)therapy
Oral PUVA was used in five generalized GA cases, resulting in clearance of
lesions in four cases within 3 to 4 months (21 to 35 treatments, cumulative exposure 100237 J/cm2) and the remaining one with perforating lesions resolved after 95 treatments
(total exposure 1859 J/cm2). 81 Recurrent lesions were present in four cases during
maintenance or when PUVA was stopped.
PUVA is useful to control GA;
paradoxically it may trigger the disease. This is somewhat akin to diseases such as
polymorphic light eruption which may both be induced and suppressed by PUVA.81
Alternatively, bath PUVA can be used.82 Both oral and bath PUVA have effect directly
on inflammatory cells or indirectly through release of anti-inflammatory mediators from
epidermal cells.80
High dose UVA-1 (daily 5 times per week for 3 weeks, starting at 60J/cm2 on day
1, 90J/cm2 on day 2 and thereafter at 130 J/cm2) was used in four generalized GA cases,83
resulting in complete clearance in one case and partial improvement in the rest. As
recurrence is possible, potential side effects of treatment should be considered,
particularly in children and young adults. PUVA should be used when other approaches
have failed and patient insists on treatment.84
9) Isotretinoin
Isotretinoin has been used in resistant generalized GA.71, 85 , 86 One case was
reported by Tang WYM et al (not included in the present study).86 Initial isotretinoin
dosage ranged from 40mg/day to 60 mg/day. A short course of treatment (3 to 4 months)
was needed in most cases, with one case of 34-year duration requiring 1 year of
treatment.85 No patients experienced significant side effects requiring premature
termination of treatment but tapering of doses was needed for deranged liver function and
The
increased lipid.71,86 Perforating GA also responded to isotretinoin. 87
immunomodulatory action of isotretinoin might be important for its effect on GA.86
Relapse in 3 out of 6 cases after several months was reported in one series,71 although
remission of up to one year was present in another.86
The present study has one case successfully managed with isotretinoin at 0.5
mg/kg/day for six months. Side effects included cheilitis that was managed with
emollient.
37
10) Dapsone
The use of dapsone was evaluated in 16 patients with GA (six localized, ten
generalized).88 Complete remission, partial remission and no response were achieved in
six, seven and one patient(s) respectively. Most response, if present, occurred within six
weeks of treatment. Relapses within three months occurred in all but one responders.
Because dapsone fails to clear the disease and carries risk of serious side effects, it should
only be used in widespread disease with caution.
Dapsone has been used in one subcutaneous GA in the present study at 50 mg/day
for six months. No clinical response is observed.
11) Other second line treatments for generalized GA
Cyclosporine was given at 4mg/kg/day for 1 month and then 2 mg/kg/day for 2
more months, which resulted in resolution of generalized GA in a case report.89 Relapse
occurred within one month after treatment was stopped. However, another generalized
GA case failed to clear with cyclosporine at 5 mg/kg/day for 2 months in the same
report.89 Systemic steroid was reported to be useful in generalized GA.72 Cyclosporin
and systemic steroid probably act through their anti-inflammatory effects.
Chlorambucil inhibits T cells and in vitro inhibits lymphocyte blastogenesis
induced by phytohemagglutinin.90 It was used at 4mg/day, which resulted in complete
healing in one within 12 weeks and partial improvement in another. Because of inherent
risk of immunosuppression and carcinogenesis, for most patients, it should not be used.
Chloroquine, 3mg/kg/day, and hydroxychloroquine, 6 mg/kg/day, were used in six
children.91 The doses were reduced to halved after three weeks and discontinued two
weeks after complete remssion. Remission was attained in all without relapse within a
mean follow up of 2.5 years. Proposed mechanisms of antimalarials include antiinflammatory effects through binding to nucleoproteins and interference with lysozyme
function. Close monitoring is mandatory, especially on the opthalmological side effects.
Potassium iodide was tested in a double blind, placebo-controlled, crossover
study involving 10 patients.92 The dosage was 3 drops increased to 10 drops three times
daily in 2 weeks for 6 months. The effect of potassium iodide over placebo was not
significant. Side effects include rhinorrhoea, metallic taste, transient acneiform eruption
and iodide goitre with or without hypothyroidism when underlying thyroid disease is
present.92, 93 Prior testing of thyroid functions is recommended before institution of
treatment. It should not be given to pregnant women.
Niacinamide, 1500mg/day for 6 months, was useful for generalized GA in a case
report. 94 This drug is a suppressor of antigen and mitogen induced lymphoblast
transformation. Potential side effects include headache, gastrointestinal upset and
hepatotoxicity. Pentoxifylline was used for GA in a case report.72 Its effect on
decreasing blood viscosity was proposed to explain the benefit, as vascular damage was
one of the postulated mechanisms of disease.
38
Vitamin E, 400 IU daily, and the 5-lipoxygenase inhibitor, zileuton 600mg four
times daily, were reported to be useful in a series of three cases.68 Remission occurred
within 3 to 4 months. Vitamin E decreases interleukin–1 beta and interleukin-18
(interleukin-18 stimulates Th1 response through induction of IFN-γ). Whereas, 5lipoxygenase inhibitor decreases production of leukotrienes which are important
inflammatory mediators. Fumaric acid esters were successful in one generalized GA
lasting 25 years.95 The exact mechanism is not known but in psoriasis, fumaric acid
esters reduce proinflammatory cytokines production and antigen presenting capacity of
monocytes as well as immunomodulate towards a Th2 response. Possible side effects
include flushing, gastrointestinal symptoms and lymphopenia.
H) Prognosis
The comparison of prognosis across various studies is difficult, as they have been
reported in a different way. In the series by Wells RS et al,23 of those who had remitted,
73% did so within 2 years and 41% of them relapsed in 1 to 8 years. In our series, of
those who have remitted, 42.9% do so within 2 years and 33.3% of them relapse in a
median of 9 months. In the series by Tan HH et al,4 67.6% of localized GA remitted after
1 year, but relapsed in 13% while their generalized GA cases did not remit after 1 to 4
years. In our series, 33% of localized GA have remission within 1 year, relapse in 19%
while 50% of generalized GA remit after a follow up of 6 to 81 months.
Besides the differences in presentation of prognosis, studies are different in terms
of follow up time and proportion of various GA subtypes and this may explain the
variation in prognosis, if any, across studies.
The present study conforms to previous series in that age and sex do not affect the
total duration of disease.3,23 Generalized GA were said to have a more protracted clinical
course than the localized subtype.25 Although a similar trend is observed in the present
study, it does not reach statistical significance.
I) Limitations of the present study
Being a case-control study, potential bias is present in the following areas:
1. Diagnostic bias
More frequent blood sugar monitoring in the case group may lead to higher
probability of DM detection in the cases as compared with controls. The present
study uses the first FBS ever checked in our clinic for comparison. However, this
does not completely solve the problem as urine sugar and spot sugar tests have
been done before in 9 cases.
2. Nonresponder
Cases still having the diseases may be more likely to attend the interview with the
author. On the other hand, subjects with a positive family history of DM may be
more likely to participate in the control group for blood sugar test.
39
3. Confounding factors for DM
While efforts have been made to minimize confounding factors in the present
study (age, sex, race-matched and skin diseases with higher DM prevalence are
excluded), the list is not exhaustive. Reported risk factors for DM are obesity,
physical inactivity, high fat diet, low birth weight and maternal gestational DM.96
It is desirable that controls have the same profile in all these confounding factors
as cases, but too much matching will result in difficulty in recruitment of enough
controls.
4. Recall bias
Some patients have defaulted follow up and their previous remissions or
exacerbations are documented by history alone that may be subjected to bias.
J) Conclusions and suggested further study
This study shows that the clinical picture and distribution of various subtypes of
GA in the local population are similar to the western literature. Age of onset of localized
GA is significantly lower than that of generalized GA. A tendency of less female cases in
local Chinese remains to be proven by a larger series. The prognosis of GA in our
population is defined, which is helpful in patient counseling and management.
Although in the present study, DM seems to be more prevalent in cases than
controls, the difference is not statistically significant. Literature review does not reveal
strong evidence in favour of DM as a causative factor for GA. But it does not exclude
the possibility of their association in other populations (where Type 1 DM is more
common) through a common autoimmune mechanism (like the association of pernicious
anemia with type 1 DM). The present study is a cross-sectional one on the prevalence of
DM in order to reduce the diagnostic bias. As a circumstantial evidence of an association
through an autoimmune mechanism, prospective follow up of GA cases and controls with
a similar schedule of DM detection will be helpful in detecting rate of DM development
over time.
OGTT is a more sensitive test then FBS alone in DM detection but it requires
more time and an additional venepuncture. As a result, recruitment of control may be
more difficult. It will be interesting to find out if there is any difference in results using
OGTT as a diabetic measure.
The way in which resolution of GA is reported in various series is different. Point
estimate of remission rate at various times is most often employed. Kaplan- Meier
estimate is theoretically better as it includes the survival experience of censored data
(data from non-remitted and defaulted cases). Standardization on reporting is suggested
to allow direct comparison of data from different studies.
40
References:
1 Cunliffe WJ. Necrobiotic disorders. In: Champion RH, Burton JL, Burns DA,
Breathnach SM, editors. Textbook of Dermatology. 6th ed. Oxford: Blackwell Science,
1998:2297-309.
2 Dahl MV. Granuloma annulare. In: Freedberg IM, Eisen AZ, Wolff K, et al, editors.
Dermatology in General Medicine. 5th ed. New York: McGraw-Hill, 1999:1152-7.
3 Studer EM, Calza AM, Saurat JH. Precipitating factors and associated diseases in 84
patients with granuloma annulare: a retrospective study. Dermatology 1996;193:364-8.
4 Tan HH, Goh CL. Granuloma annulare: a review of 41 cases at the national skin center.
Ann Acad Med Singapore 2000;29:714-8.
5 Smith MD, Downie JB, DiCostanzo D. Granuloma annulare. Int J Dermatol
1997;36:326-33.
6 Rhodes EL, Hill DM, Ames AC, Tourle CA, Taylor CG. Granuloma annulare.
Prednisone glycosuria tests in a non-diabetic group. Br J Dermatol 1966;78:532-5.
7 Haim S, Friedman-Birnbaum R, Shafrir A. Generalized granuloma annulare:
relationship to diabetes mellitus as revealed in 8 cases. Br J Dermatol 1970;83:302-5.
8 Hammond R, Dyess K, Castro A. Insulin production and glucose tolerance in patients
with granuloma annulare. Br J Dermatol 1972;87:540-7.
9 Haim S, Friedman-Birnbaum R, Haim N, Shafrir A, Ravina A. Carbohydrate tolerance
in patients with granuloma annulare- study of fifty-two cases. Br J Dermatol
1973;88:447-51.
10 Muhlemann MF, Williams DRR. Localized granuloma annulare is associated with
insulin-dependent diabetes mellitus. Br J Dermatol 1984;111:325-9.
11 Kidd GS, Graff GE, Davies BF, et al. Glucose tolerance in granuloma annulare.
Diabetes Care 1985;8:380-4.
12 Shimizu H, Harada T, Baba E, Kuramochi M, et al. Perforating granuloma annulare.
Int J Dermatol 1985;24:581-3.
13 Veraldi S, Bencini PL, Drudi E, Caputo R. Laboratory abnormalities in granuloma
annulare: a case-control study. Br J Dermatol 1997;136:652-3.
14 Gannon TF, Lynch PJ. Absence of carbohydrate intolerance in granuloma annulare. J
Am Acad Dermatol 1994:30:662-3.
41
15 Nebesio CL, Lewis C, Chuang TY. Lack of association between granuloma annulare
and type 2 diabetes mellitus. Br J Dermatol 2002;146:122-4.
16 Perez MI, Kohn SR. The cutaneous manifestation of diabetes mellitus. J Am Acad
Dermatol 1994; 30:519-31.
17 The expert committee on the diagnosis and classification of diabetes mellitus. Report
on the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes
Care 2002; 25 supplement 1:s5-20.
18 Alberti KGMM, Zimmet PZ for the WHO Consultation. Definition, Diagnosis and
Classification of Diabetes Mellitus and its complications. Part 1: Diagnosis and
Classification of Diabetes Mellitus. Provisional Report of a WHO consultation. Diabet
Med 1998; 15:539-53.
19 Janus ED, Wat NM, Lam KS, et al. The prevalence of diabetes, association with
cardiovascular risk factors and implications of diagnostic criteria (ADA 1997 and WHO
1998) in a 1996 community-based population study in Hong Kong Chinese. Hong Kong
Cardiovascular Risk Factor Steering Committee. American Diabetes Association. Diabet
Med. 2000 ;17:741-5.
20 Stern MP, Burke JP. Impaired glucose tolerance and impaired fasting glucose: risk
factors or diagnostic categories. In: LeRoith D, Taylor SI, Olefsky JM, editors. Diabetes
mellitus a fundamental and clinical text. 2nd edition. Philadelphia: Lippincott Williams
and Wilkins, 2000:558-66.
21 Janus ED for the cardiovascular risk factor prevalence study group. The Hong Kong
cardiovascular risk prevalence study, 1995-1996. Hong Kong: Department of Clinical
Biochemistry, Queen Mary Hospital, 1997:47-53.
22 Sweetman SS. Corticosteroids. In: Martindale: the complete drug reference. 33rd ed.
London: Pharmaceutical Press, 2002:1039-81.
23 Wells RS, Smith MA. The natural history of granuloma annulare. Br J Dermatol
1963;75:199-205.
24 Muhlbauer JE. Granuloma annulare. J Am Acad Dermatol 1980;3:217-30.
25 Dabski K, Winkelmann RK. Generalized granuloma annulare: clinical and laboratory
findings in 100 patients. J Am Acad Dermatol 1989;20:39-47.
26 Friedman-Birnbaum R. Generalized and localized granuloma annulare. Int J Dermatol
1986;25:364-6.
42
27 Penas PF, Jones-Caballero M, Fraga J, Sanchez-Perez J, Garcia-Diez A. Perforating
granuloma annulare. Int J Dermatol 1997;30:340-8.
28 Abrusci V, Weiss E, Planas G. Familial generalized perforating granuloma annulare.
Int J Dermatol 1988;27:126-7.
29 Weedon D. The granulomatous reaction pattern. Skin Pathology. 2nd ed. London:
Churchill Livingstone, 2002:194-220.
30 Lucky AW, Prose NS, Bove K, White WL, Jorizzo JL. Papular umbilicated granuloma
annulare a report of four paediatric cases. Arch Dermatol 1992;128:1375-8.
31 Mutasim DF, Bridges AG. Patch granuloma annulare: clinicopathologic study of 6
patients. J Am Acad Dermatol 2000;42:417-21.
32 Vargas-Diez E, Feal-Cortizas C, Fraga J, Fernandez-Herrera J, Garcia-Diez A.
Follicular pustulous granuloma annulare. Br J Dermatol 1998;138:1075-8.
33 Harpster EF, Mauro T, Barr RJ. Linear granuloma annulare. J Am Acad Dermatol
1989;21:1138-41.
34 Dahl MV. Is actinic granuloma really granuloma annulare. Arch Dermatol
1986;122:39-40.
35 Ackerman AB, Chongchitnant N, Sanchez J, et al. Histologic diagnosis of
inflammatory skin diseases. 2nd ed. Baltimore: William and Wilkins, 1997:411-2.
36 Hanke CW, Bailin PL, Roenigk HH. Annular elastolytic giant cell granuloma. J Am
Acad Dermatol 1979;1:413-21.
37 Felner EI, Steinberg JB, Weinberg AG. Subcutaneous granuloma annulare: a review
of 47 cases. Pediatrics 1997;100:965-7.
38 Umbert P, Winkelmann RK. Histologic, ultrastructural and histochemical studies of
granuloma annulare. Arch Dermatol 1977;113:1681-6.
39 Dabski K, Winkelmann RK. Generalized granuloma annulare: histopathology and
immunopathology. J Am Acad Dermatol 1989;20:28-39.
40 Adams BB, Mutasim DF. Co-localization of granuloma annulare and mid-dermal
elastolysis. J Am Acad Dermatol 2003;48:S25-7.
41 Ozkan S, Fetil E, Izler F, et al. Anetoderma secondary to generalized granuloma
annulare. J Am Acad Dermatol 2000;42:335-8.
43
42 Dahl MV, Ullman S, Goltz RW. Vasculitis in granuloma annulare. Arch Dermatol
1977;113:463-7.
43 Umbert P, Winkelmann RK. Granuloma annulare: direct immunofluorescence study.
Br J Dermatol 1976;95:487-92.
44 Bergman R, Pam Z, Lichtig C, Reiter I, Friedman-Birnbaum R. Localized granuloma
annulare: histopathological and direct immunofluorescence study of early lesions, and the
adjacent normal-looking skin of actively spreading lesions. Am J Dermatopath
1993;15:544-8.
45 Friedman-Birnbaum R, Gideoni O, Bergman R, Pollack S. A study of HLA antigen
association in localized and generalized granuloma annulare. Br J Dermatol
1986;115:329-33.
46 Gibney MD, Nahass GT, Leonardi CL. Cutaneous reactions following herpes zoster
infections: report of three cases and a review of the literature. Br J Dermatol
1996;134:504-9.
47 Requena L, Kutzner H, Escalonilla P, Ortiz S, Schaller J, Rohwedder A. Cutaneous
reactions at sites of herpes zoster scars: an expanded spectrum. Br J Dermatol
1998;138:161-8.
48 Spencer SA, Fenske NA, Espinoza CG, Hamill JR, Cohen LE, Espinoza LR.
Granuloma annulare-like eruption due to chronic Esptein-Barr virus infection. Arch
Dermatol 1988;124:250-5.
49 Toro JR, Chu P, Yen TSB, LeBoit PE. Granuloma annulare and human
immunodeficiency virus infection. Arch Dermatol 1999;135:1341-6.
50 Granel B, Serratrice J, Rey J, et al. Chronic hepatitis C virus infection associated with
a generalized granuloma annulare. J Am Acad Dermatol 2000;43:918-9.
51 O’Moore EJ, Nandawni R, Uthayakumar S, Nayagam AT, Darley CR. HIV-associated
granuloma annulare: a report of six cases. Br J Dermatol 2000;142:1054-5.
52 Winkelmann RK. The granuloma annulare phenotype and tuberculosis. J Am Acad
Dermatol 2002;46:948-52.
53 Aberer E, Schmidt BL, Breier F, Kinaciyan T, Luger A. Amplification of DNA of
Borrelia burgdorferi in urine samples of patients with granuloma annulare and lichen
sclerosus et atrophicus. Arch Dermatol 1999;135:210-2.
44
54 Vazquez-Lopez F, Pereiro M Jr, Manjon Haces JA. Localized granuloma annulare and
autoimmune thyroiditis in adult women: a case-control study. J Am Acad Dermatol
2003;48:517-20.
55 Umbert P, Winkelmann RK. Granuloma annulare and sarcoidosis. Br J Dermatol
1977;97:481-6.
56 Antony F, Holden CA. Sweet’s syndrome in association with generalized granuloma
annulare in a patient with previous breast carcinoma. Clin Exp Dermatol 2001;26:668-70.
57 Ben-Amitai D, Hodak E, Lapidoth M, David M. Co-existing morphoea and
granuloma annulare- are the conditions related? Clin Exp Dermatol 1999;24:86-9.
58 Fukai K, Ishii M, Kobayashi H, Someda Y, Hamada T, Tsujino S. Generalized
granuloma annulare in a patient with temporal arteritis- are these conditions associated?
Clin Exp Dermatol 1990;15: 70-2.
59 Li A, Hogan DJ, Sanusi ID, Smoller BR. Granuloma annulare and malignant
neoplasm. Am J Dermatopath 2003;25:113-6.
60 Schlesselman JJ, Stolley PD. Research strategies. In: Schlesselman JJ editor. Casecontrol studies, design, conduct, analysis. New York: Oxford University Press, 1982:7-26.
61 Majewski BBJ, Koh MS, Barter S, Rhodes EL. Increased factor VIII-related antigen
in necrobiosis lipoidica and widespread granuloma annulare without associated diabetes.
Br J Dermatol 1982;107:641-5.
62 Baykal C, Ozkaya-Bayazit E, Kaymaz R. Granuloma annulare possibly triggered by
antitetanus vaccination. J Eur Acad Dermatol Venereol 2002;16:516-8.
63 Lim AC, Hart K, Murrell D. A granuloma annulare-like eruption associated with the
use of amlodipine. Australas J Dermatol 2002;45:24-7.
64 Buechner SA, Winkelmann RK, Banks PM. Identification of T-cell subpopulations in
granuloma annulare. Arch Dermatol 1983;19:125-8.
65 Umbert P, Belcher RW, Winkelmann RK. Lymphokines (MIF) in the serum of
patients with sarcoidosis and cutaneous granuloma annulare. Br J Dermatol 1976;95:4815.
66 Cherney KJ, Lindroos WE, Goltz RW, Dahl MV. Leucocyte function in granuloma
annulare. Br J Dermatol 1979;101:23-31
45
67 Mempel M, Musette P, Flageul B, et al. T-cell receptor repertoire and cytokine pattern
in granuloma annulare: defining a particular type of cutaneous granulomatous
inflammation. J Invest Dermatol 2002;118:957-66.
68 Smith KJ, Norwood C, Skelton H. Treatment of disseminated granuloma annulare
with a 5-lipoxygenase inhibitor and vitamin E. Br J Dermatol 2002;146:667-70.
69 Kallioinen M, Sandberg M, Kinnunen T, Oikarinen A. Collagen synthesis in
granuloma annulare. J Invest Dermatol 1992;98:463-8.
70 Saarialho-Kere UK, Chang ES, Welgus HG, Parks WC. Expression of interstitial
collagenase, 92-kDa gelatinase and tissue inhibitor of metalloproteinases-1 in granuloma
annulare and necrobiosis lipoidica diabeticorum. J Invest Dermatol 1993;100:335-42.
71 Schleicher SM, Milstein HJ, Lim SJM, Stanton CD. Resolution of disseminated
granuloma annulare with isotretinoin. Int J Dermatol 1992;31:371-2.
72 Ilchyshyn A. Granuloma annulare. In: Lebwohl MG, Heymann WR, Berth-Jones J,
Coulson I, editors. Treatment of skin disease comprehensive therapeutic strategies.
London: Mosby, 2002: 247-50.
73 Levin NA, Patterson JW, Yao LL, Wilson BB. Resolution of patch-type granuloma
annulare lesions after biopsy. J Am Acad Dermatol 2002;46:426-9.
74 Sparrow G, Abell E. Granuloma annulare and necrobiosis lipoidica treated by jet
injector. Br J Dermatol 1975;93:85-9.
75 Blume-Peytavi U, Zouboulis CC, Jacobi H, Scholz A, Bisson S, Orfanos CE.
Successful outcome of cryosurgery in patients with granuloma annulare. Br J Dermatol
1994;130:494-7.
76 Weiss J, Muchenberger S, Schopf E, Simon JC. Treatment of granuloma annulare by
local injections with low-dose recombinant human interferon gamma. J Am Acad
Dermatol 1998;39:117-9.
77 Wilkin JK, Ducomb D, Castrow FF. Scarification treatment of granuloma annulare.
Arch Dermatol 1982;118:68-9.
78 Shelley WB, Shelley ED. Surgical pearl: Surgical treatment of tumor-sized granuloma
annulare of the fingers. J Am Acad Dermatol 1997;37:473-4.
79 Kuwahara RT, Skinner RB Jr. Granuloma annulare resolved with topical application
of imiquimod. Pediatr Dermatol 2002;19:368-9.
46
80 Grundmann-Kollmann M, Ochsendorf FR, Zollner TM, Tegeder I, Kaufmann R,
Podda M. Cream psoralen plus ultraviolet A therapy for granuloma annulare. Br J
Dermatol 2001;144:996-9.
81 Kerker BJ, Huang CP, Morison WL. Photochemotherapy of generalized granuloma
annulare. Arch Dermatol 1990;126:359-61.
82 Szegedi A, Begany A, Hunyadi J. Successful treatment of generalized granuloma
annulare with polyethylene sheet bath PUVA. Acta Derm Venereol 1999;79:84-5.
83 Muchenberger S, Schopf E, Simon JC. Phototherapy with UVA-1 for generalized
granuloma annulare. Arch Dermatol 1997;133:1605.
84 Schwarz T, Rutter A, Hawk J. Phototherapy and photochemotherapy: less common
indications for its use. In: Krutmann, Honigsmann, Elmets CA, Bergstresser PR, editors.
Dermatological phototherapy and photodiagnostic methods. Berlin :Springer 2001:17997.
85 Adams DC, Pa D, Hogan DJ. Improvement of chronic generalized granuloma
annulare with isotretinoin. Arch Dermatol 2002;138:1518-9.
86 Tang WYM, Chong LY, Lo KK. Resolution of generalized granuloma annulare with
isotretinoin therapy. Int J Dermatol 1996;35:455-6.
87 Ratnavel RC, Norris PG. Perforating granuloma annulare: response to treatment with
isotretinoin. J Am Acad Dermatol 1995;32:126-7.
88 Steiner A, Pehamberger H, Wolff K. Sulfone treatment of granuloma annulare. J Am
Acad Dermatol 1985;13:1004-8.
89 Ho VC. Cyclosporine in the treatment of generalized granuloma annulare. J Am Acad
Dermatol 1995;32:298.
90 Kossard S, Winkelmann RK. Response of generalized granuloma annulare to
alkylating agents. Arch Dermatol 1978;114:216-220.
91 Simon M, von den Driesch P. Antimalarials for control of disseminated granuloma
annulare in children. J Am Acad Dermatol 1994;31:1064-5.
92 Smith JB, Hansen D, Zone JJ. Potassium iodide in the treatment of disseminated
granuloma annulare. J Am Acad Dermatol 1994;30:791-2.
93 Caserio RJ, Eaglstein WH, Allen CM. Treatment of granuloma annulare with
potassium iodide. J Am Acad Dermatol 1984;10:294.
47
94 Ma A, Medenica M. Response of generalized granuloma annulare to high dose
niacinamide. Arch Dermatol 1983;119:836-9.
95 Kreuter A, Gambichler T, Altmeyer P, Brockmeyer NH. Treatment of disseminated
granuloma annulare with fumaric acid esters. BMC Dermatol 2002;2:5.
96 Bennett PH. Epidemiology of Type 2 diabetes mellitus. In: LeRoith D, Taylor SI,
Olefsky JM, editors. Diabetes mellitus a fundamental and clinical text. 2nd edition.
Philadelphia: Lippincott Williams and Wilkins, 2000:544-8.
48
Acknowledgement
The author wishes to take this opportunity to thank Dr. Lo Kuen Kong, my
present program director, and Dr. Chong Lai Yin who have given enormous supervision
and guidance throughout the training program. Dr. Leung Chi Yan, my former program
director, has provided me with inspiration during my budding stage as a training
dermatologist and venereologist. My supervisor, Dr. Au Tak Shing, has given me
invaluable advice on the dissertation ever since in the planning stage. He has kindly
reviewed this manuscript and has offered professional opinion on the text and statistical
analysis.
The study is made possible and interesting with the support from the four senior
pathologists: Drs Lam Wing Yin (Tuen Mun Hospital), Lee Kam Cheong (Princess
Margaret Hospital), Lee King Chung (Queen Elizabeth Hospital) and Yau Ko Chiu
(Pathology Institute, Department of Health).
I would like to thank Drs. Au Tak Shing, Chan Yui Hoi, Ho Hing Fung, Ho Ka
Keung, Ho Man Hon, Hui Shiu Kee, Ku Lap Shing, Tang Wai Ki and Yeung Kwok Hung
for helping me in recruiting control subjects in various clinics.
The nurses in the Social Hygiene Service have helped me to retrieve the old
clinical photograph and perform the blood taking. Last but not least, I want to thank the
granuloma annulare patients and the control subjects who have kindly agreed to
participate in the study.
49
Appendix 1
Granuloma annulare case data
Demographic data:
Clinic:____________
Patient no.:______
Name:_______________
Sex: M/F
Age:____
Clinic no:____________
Ethnic group:__________
Clinical Data:
Onset of lesion
Morphology of lesion:
Age:______
Date:________
Annular/ papular/ perforating/
Others/ Mixed:
subcutaneous
Distribution:
Arm:___
Elbow:___
Forearm:____
Wrist:___
Hands(D/P):____
Thigh:___
Knee:___
Calf/Shin:___
Ankle:___
Feet(D/S):____
Chest:____
Abdomen:___
Back:____
Face:____
Scalp:____
Ears:____
Genitalia:____
Others:
Total no. of lesions: ___
Generalized or Localized
Symptoms: pruritic/ pain/others:
? in remission now Y/N
If yes, duration of disease:_____________
? Ever remitted Y/N
Recurrence:
______________________________________________________________
______________________________________________________________
______________________________________________________________
Precipitating Factors
Past Health:
Sunburn
Drug
URTI
Insect bite
Weather
Pregnancy
Trauma
Stress
Others
DM: Y/N for ______ yrs
If DM+ve on insulin Y/N
DM diagnosed by FBS/spot sugar/OGTT
Others:
1.__________________
2.________________
3.__________________
4.________________
5.__________________
6.________________
Family History:
Agree for blood test
Previous DM screen
Treatment:
1. GA Y/N
2. DM Y/N
Y/N
FBS:____ (
Type
Duration
Y/N
/ / )
Method: urine sugar/FBS/spot sugar/OGTT
Response
1.
2.
3.
50
Appendix 1
51
Appendix 2
Clinic:
Date: ________________
Data for Granuloma annulare control cases
1. Year of birth
_________
2. Sex
__M / F__
3. Family history of DM
__Y / N__
4. Agree for study (0=disagree, 1=agree)
___________
If agree for study, proceed to the following sessions:
5.
Personal history of DM
a) ? DM (0= -ve, 1= +ve)
___________
b) Method used if (a) is +ve (0=urine sugar, 1= spot sugar, 2= FBS, 3= OGTT) __________
________
c) If DM +ve, on oral drugs or insulin? (0= OHA, 1=insulin)
6. Dermatology diagnosis
_________________________________
7. Clinic no.: _______________
8. FBS: _________
51
Appendix 3
Memo to Granuloma Annulare Clients
Granuloma annulare: review of cases in the Social Hygiene Service and a casecontrol study of its association with diabetes mellitus
We are conducting a study of granuloma annulare. The aetiology of the disease is still
not known and its relationship with diabetes mellitus is inconclusive. Some studies
showed a relationship of the disease with diabetes mellitus while others did not. Most
of the studies were carried out in the Western population. We are collecting
information from clients with granuloma annulare in the Social Hygiene Service. The
required data include age of onset, sex, region of involvement, response to therapy,
recurrence history and past medical illness of the clients especially the diabetic status,
etc. If you are not known to have diabetes mellitus in the past, we would like to invite
you to have a fasting blood sugar test. All information collected is confidential. If
there is any abnormality detected in the fasting blood sugar, we shall refer you to the
government general outpatient clinic or the specialist clinic for follow up.
You are reminded that participation in the study is entirely voluntary. There
will be no difference in the quality of care you will receive in this clinic whether you
agree to participate in the study or not.
Social Hygiene Service, Department of Health
環狀肉芽腫病人備忘
環狀肉芽腫：病歷回顧及與糖尿病關連之對照研究
我們正從事皮膚病環狀肉芽腫的研究。這種皮膚病的成因仍然不明。有部份的
研究顯示此病與糖尿病有關連，但亦有研究否定這一點，現在還沒有結論，而
且大部份的研究是在外國進行。我們希望集合在香港社會衛生科內證實患有這
種皮膚病患者進行研究；範圍包括發病年齡、性別、發病範圍、對治療之反
應、有否復發等及與其他病患尤其是與糖尿病之關係。若果你沒有糖尿病，我
們會邀請閣下空腹檢驗血糖。所有個人資料，絕對保密。若血糖出現不正常的
情況，我們會轉介你往其他政府或專科門診進行跟進。
你是否參加這項研究，純屬自願，你不會因為拒絕參加而受到次等待遇。
衛生署社會衛生科
52
Appendix 4
Memo to control clients
Granuloma annulare: review of cases in the Social Hygiene Service and a casecontrol study of its association with diabetes mellitus
We are now conducting a study to assess the relationship between granuloma
annulare (a skin disease) and diabetes mellitus. You have been chosen as a subject in
the control group. We would like to collect data on the diagnosis of your skin
problem, and personal and family history of diabetes mellitus for analysis. If you are
not known to have diabetes mellitus, we would like to invite you to have a fasting
blood sugar test. All information collected is confidential. If there is any abnormality
detected in the fasting blood sugar, we shall refer you to the government general
outpatient clinic or the specialist clinic for follow up.
You are reminded that participation in the study is entirely voluntary. There
will be no difference in the quality of care you will receive in this clinic whether you
agree to participate in the study or not.
Social Hygiene Service, Department of Health
對照病人備忘
環狀肉芽腫：病歷回顧及與糖尿病關係之對照研究
我們正在進行關於環狀肉芽腫 (一種皮膚病) 與糖尿病關係之研究。閣下已被抽
選為對照之組別以作比較。所需要的資料包括閣下之皮膚病診斷，本人及家族
糖尿病的病史。若果閣下沒有糖尿病，我們將誠邀閣下空肚抽驗血糖。所有個
人資料，絕對保密。如發現血糖有不正常的情況，我們將轉介閣下往政府門診
或專科進行跟進。
你是否參加這項研究，純屬自願，你不會因為拒絕參加而受到次等待遇。
衛生署社會衛生科
53
Appendix 5
Definitions of terms
Age of onset
Age when skin lesions were first noticed
Family history of DM
Family history of DM in first degree relative
Granuloma
An accumulation of histiocytes
Palisading granuloma
Histiocytic infiltration surrounding altered collagen,
with the long axis of cells perpendicular to the
necrobiotic areas
Interstitial granuloma
histiocytes infiltrating between and about the collagen
fibres
Sarcoidal granuloma
Aggregations of epitheloid histiocytes, occasional
giant cells and a rim of lymphocytes that resemble
sarcoidosis
Granuloma annulare
A degenerative disease of the skin, characterized by
focal degeneration of collagen with surrounding areas
of reactive inflammation and fibrosis
Mucin
Appears as faint blue material with a stringy and
granular appearance on routine hematoxylin and eosin
stain, can be highlighted by alcian blue stain or
colloidal iron stain
Necrobiosis
Refers to non-specific connective tissue alteration,
usually associated with various granulomatous
inflammation of skin
Time for complete remission Interval from onset of disease to complete remission
Time for relapse
Interval from complete remission to recurrence of
disease
54
Appendix 5
Treatment response
Owing to the retrospective nature of study, partial
remission is difficult to assess and response is divided
into persistent disease and complete remission
Complete remission
Disappearance of all skin lesions, sometimes with
residual hyperpigmentation or scarring
Persistent disease
Evidence of active skin disease of various severity
Relapse
Recurrence of skin lesions after complete remission
55
Plate 1. Localized GA with skin coloured papules over the left elbow. A
vague arciform pattern is noted in the centre. There is no epidermal change.
Plate 2. Typical GA with erythematous nonscaly arciform lesions over the
left elbow.
56
Plate 3. GA with annular lesions over both hand dorsa and papular lesions
over the distal forearms. Edge of skin lesions is raised.
Plate 4. GA: predominant papular form with slightly erythematous papules
over the distal forearms.
57
Plate 5. Generalized GA: coalescing papules without central clearing,
forming plaques over both elbows.
Plate 6. Generalized GA with well defined erythematous plaques over the
abdominal wall. Surface changes are absent (same patient as Plate 5).
58
Plate 7 a and b. Generalized GA with extensive coalescing papules over the
chest, abdomen, back and arms.
59
Plate 8. Generalized GA with annular lesions over the abdominal wall.
Plate 9. Generalized GA with multiple lesions over the right elbow. Some of
the lesions have fused to give a serpiginous outline (same patient as Plate 8).
60
Plate 10. Generalized GA, same patient as Plate 8. Central clearing is less
evident in the right shin and foot dorsum giving a plaque appearance.
Plate 11. Generalized GA with annular and serpiginous lesion over the right
abdominal wall and right hip. Postinflammatory pigmentation over the
centre of lesions can sometimes occur.
61
Plate 12. Perforating GA with skin colour papules over the hand dorsum.
Umbilication is present in papules over dorsum of the third
metacarpophalangeal joint.
Plate 13. Histology: typical palisading granuloma with central necrobiosis.
The collagen fibres are more fragmented and separated from each other.
Mucin, which appears as bluish stringy and granular material between
collagen fibres, can sometimes be noted on hematoxylin and eosin stain.
62
Plate 14. Histology: interstitial pattern showing histiocytes, with large and
occasional reniform nuclei and prominent cytoplasm, infiltrate between
collagen fibres. Necrobiosis is less prominent in this form of GA.
Plate 15. Histology: histiocytes and occasional giant cells are the main
inflammatory cells in GA.
63
Plate 16. Histology: nuclear dust can sometimes be identified in GA lesions.
They may represent apoptotic inflammatory cells or other cell types.
Plate 17. Histology: perforating GA with a ‘busy looking’ superficial dermis.
Necrobiosis, which appears as pinkish material, is already evident in this
scanning view. A central ‘dip’ is also present, which corresponds clinically
to the umbilicated papule.
64
Plate 18. Histology: perforating GA (higher power of Plate 17). Note the
palisading granuloma, epidermal destruction and impending rupture of
epidermis.
65

Granuloma annulare: review of cases in the Social Hygiene

Transcription

Similar documents

Disseminated granuloma annulare - RJME

Case Report Fractional Photothermolysis for the Treatment of

Noninfectious Granulomas

What Kind of Theory Is a Model-‐Theore:c Seman:cs of a Natural

Buschke-Ollendorff syndrome accidentally diagnosed after a left

Spectrum of imaging features of SAPHO/ CRMO syndrome

Intestinal anisakiasis in Italy: a case treated by emergency surgery

experienced. focused. evolving. just like you.

Endoscopic Tattoo of the Colon might be Standardized to

ONYCHOLYSIS DUE TO TRAUMA Patricia Chang1, Monica Vanesa

Mast cell activation disorders: chronic urticaria

WCET Journal

Acne: A Holistic Approach